THE PSYCHOLOGY OF GROUP PERCEPTION
THE PSYCHOLOGY OF GROUP PERCEPTION Perceived Variability, Entitativity, and Essentialism Edited by
Vincent Yzerbyt Charles M.Judd and Olivier Corneille
Psychology Press New York and Hove
Published in 2004 by Psychology Press 29 West 35th Street New York, NY 10001 www.psypress.com Published in Great Britain by Psychology Press 27 Church Road Hove, East Sussex BN3 2FA www.psypress.co.uk Copyright © 2004 by Taylor & Francis, Inc. Psychology Press is an imprint of the Taylor & Francis Group. This edition published in the Taylor & Francis e-Library, 2005. “To purchase your own copy of this or any of Taylor & Francis or Routledge’s collection of thousands of eBooks please go to www.eBookstore.tandf.co.uk.” All rights reserved. No part of this book may be reprinted or reproduced or utilized in any form or by any electronic, mechanical or other means, now known or hereafter invented, including photocopying and recording or in any information storage or retrieval system, without permission in writing from the publisher. www.socialpsychologyarena.com Library of Congress Cataloging-in-Publication Data The psychology of group perception: perceived variability, entitativity, and essentialism/ Vincent Yzerbyt, Charles M.Judd & Olivier Corneille, ed[itor]s. p. cm. Includes bibliographical references and index. ISBN 1-84169-061-9 (Print Edition) 1. Social groups. I. Yzerbyt, Vincent. II. Judd, Charles M. III. Corneille, Olivier. HM716.P79 2003 302.3–dc21 2003010696
ISBN 0-203-64497-2 Master e-book ISBN
ISBN 0-203-67690-4 (Adobe eReader Format)
Contents
1
About the Editors
vii
Contributors
viii
Preface
xii
Perceived Variability, Entitativity, and Essentialism: Introduction and Overview Vincent Yzerbyt, Charles M.Judd and Olivier Corneille
1
SECTION 1. CONSTRUCTS AND DEFINITIONS 2
Dynamic Entitativity: Perceiving Groups as Actors Marilynn B.Brewer, Ying-Yi Hong and Qiong Li
19
3
Perceiving the Groupness of Groups: Entitativity, Homogeneity, Essentialism, and Stereotypes David L.Hamilton, Steven J. Sherman and Julie S.Rodgers
30
4
Essentialism and Entitativity: Structures of Beliefs about the Ontology of Social Categories Nick Haslam, Louis Rothschild and Donald Ernst
47
5
The Mental Representation of Social Categories: Category Boundaries, Entitativity, and Stereotype Change Myron Rothbart and Bernadette Park
60
6
Subjective Essentialism in Action: Self-Anchoring and Social Control as Consequences of Fundamental Social Divides Vincent Yzerbyt, Claudia Estrada, Olivier Corneille, Eléonore Seron and Stephanie Demoulin
77
SECTION 2. ANTECEDENTS AND CONSEQUENCES 7
In the Eye of the Beholder: Lay Theories and the Perception of Group Entitativity, Variability, and Essence Jason E.Plaks, Sheri R.Levy, Carol S.Dweck and Steven J.Stroessner
96
8
Components of Homogeneity: A Multiple-Process Model of Social Categorization Karl Christoph Klauer, Katja Ehrenberg and Ingo Wegener
110
9
Forming Stereotypes of Entitative Groups
121
v
Craig McGarty 10
From Basketball to Business: Expertise, Implicit Covariation, and Social Judgment Patricia W.Linville and Gregory W.Fischer
135
11
Gender Outgroup Homogeneity: The Roles of Differential Familiarity, Gender Differences, and Group Size Mark Rubin Miles Hewstone, Richard J.Crisp, Alberto Voci and Zoë Richards
153
12
Group Size, Outcome Dependency, and Power: Effects on Perceived and Objective Group Variability Ana Guinote
166
13
The Acquisition, Transmission, and Discussion of Social Stereotypes: Influences of Communication on Group Perceptions Markus Brauer, Charles M.Judd and Micah S.Thompson
178
14
Culture, Communication, and Entitativity: A Social Psychological Investigation of Social Reality Yoshihisa Kashima
193
15
Group Socialization, Uncertainty Reduction, and the Development of New Members’ Perceptions of Group Variability Carey S.Ryan,Debbie R.Robinson and Leslie R.M.Hausmann
206
16
Entitativity, Group Distinctiveness, and Social Identity: Getting and Using Social 219 Structure Russell Spears, Daan Scheepers, Jolanda Jetten, Bertjan Doosje, Naomi Ellemers and Tom Postmes
17
Social Identity as the Basis of Group Entitativity: Elaborating the Case for the “Science of Social Groups Per Se” Katherine J.Reynolds, Penelope J.Oakes, S.Alexander Haslam, John C.Turner and Michelle K.Ryan
236
18
The Perception of Outgroup Threat: Content and Activation of the Outgroup Schema Tim Wildschut, Chester A.Insko and Brad Pinter
249
19
The Maintenance of Entitativity: A Subjective Group Dynamics Approach Dominic Abrams, José M.Marques, Georgina Randsley de Moura, Paul Hutchison and Nicola J.Bown
268
20
On the Advantages of Reifying the Ingroup Emanuele Castano
283
21
Uncertainty and Extremism: Identification with High Entitativity Groups under Conditions of Uncertainty Michael A.Hogg
298
22
Entitativity and Social Integration: Managing Beliefs about the Reality of Groups Richard L.Moreland and Jamie G.McMinn
311
vi
References
326
Index
371
About the Editors
Vincent Yzerbyt is a professor of psychology at the Catholic University of Louvain at Louvain-la-Neuve, Belgium. Dr. Yzerbyt received his Ph.D. from the University of Louvain in 1990 and has written and edited several books dealing with social cognition as well as journal articles on intergroup relations and stereotyping. He served as Associate Editor of the British Journal of Social Psychology and the Personality and Social Psychology Bulletin. Dr. Yzerbyt is currently the President of the European Association of Experimental Social Psychology. Charles M.Judd is a professor of psychology at the University of Colorado, Boulder. Dr. Judd received his Ph.D. from Columbia University in 1976, and was an assistant professor of psychology at Harvard University until 1981. He is the former editor of the Journal of Experimental Social Psychology, as well as the associate editor of the Journal of Personality and Social Psychology and the Psychological Review. Olivier Corneille graduated from the University of Liège, Belgium in 1992 and obtained his doctoral degree in psychology from the Catholic University of Louvain in 1996. He is currently a Research Associate at the Belgian National Fund for Scientific Research. His research is mainly concerned with the impact of categorization on social perception and behavior.
Contributors
Dominic Abrams University of Kent at Canterbury Nicola J.Bown University of Leeds Markus Brauer CNRS, Université Blaise Pascal Marilynn B.Brewer Ohio State University Emanuele Castano Graduate Faculty, New School University Olivier Corneille Catholic University of Louvain at Louvain-la-Neuve Richard J.Crisp University of Birmingham Stéphanie Demoulin Catholic University of Louvain at Louvain-la-Neuve Bertjan Doosje University of Amsterdam Carol S.Dweck Columbia University Katja Ehrenberg Universität Bonn
ix
Naomi Ellemers Leiden University Donald Ernst Hillsdale College Claudia Estrada Catholic University of Louvain at Louvain-la-Neuve Gregory W.Fischer Duke University Ana Guinote University of Kent David L.Hamilton University of California, Santa Barbara S.Alexander Haslam University of Exeter Nick Haslam University of Melbourne Leslie R.M.Hausmann University of Pittsburgh Miles Hewstone University of Oxford Michael A.Hogg University of Queensland Ying-Yi Hong University of Illinois Paul Hutchison University of Kent at Canterbury Chester A.Insko University of North Carolina, Chapel Hill Jolanda Jetten University of Exeter Charles M.Judd University of Colorado, Boulder Yoshihisa Kashima University of Melbourne Karl Christoph Klauer Universität Bonn Sheri T.Levy State University of New York at Stony Brook Qiong Li University of Maryland Patricia W.Linville Duke University
x
José M.Marques Univerity of Porto Craig McGarty The Australian National University Jamie G.McMinn University of Pittsburgh Richard L.Moreland University of Pittsburgh Penelope J.Oakes The Australian National University Bernadette Park University of Colorado, Boulder Brad Pinter University of Washington Jason E.Plaks University of Washington Tom Postmes University of Exeter Georgina Randsley de Moura University of Kent at Canterbury Katherine J.Reynolds The Australian National University Zoë Richards Cardiff University Debbie R.Robinson University of Pittsburgh Julie S.Rodgers University of California, Santa Barbara Myron Rothbart University of Oregon Louis Rothschild Brown University Mark Rubin University of Newcastle, Australia Carey S.Ryan University of Nebraska at Omaha Michelle K.Ryan The Australian National University Daan Scheepers Leiden University Eléonore Seron Catholic University of Louvain at Louvain-la-Neuve
xi
Steven J.Sherman Indiana University Russell Spears Cardiff University Steve J.Stroessner Barnard College Micah S.Thompson University of Colorado, Boulder John C.Turner The Australian National University Alberto Voci University of Padova, Italy Ingo Wegener Universität Bonn Tim Wildschut University of Southampton Vincent Yzerbyt Catholic University of Louvain at Louvain-la-Neuve
Preface
This book emerged out of the enthusiasm that followed a meeting held in Louvain-la-Neuve, Belguim in July 1999. The meeting, devoted to the topic of group perception and cognition, was organized under the auspices of the European Association of Experimental Social Psychology. By all indications the meeting was a huge success. We were fortunate to be able to bring together 25 leading researchers from all over the world. All 25 presented their latest work on the important issues of group perception, and these presentations led us to focus on the particular issues surrounding the central themes of this book: Entitativity, Perceived Variability, and Essentialism of social groups. We realized during the course of the meeting that these central themes were closely intertwined, that there were very different schools of thought about these constructs, and that the interchange that occurred at the meeting and beyond might fruitfully advance thinking about these issues as central to our understanding of group perception. This is not to say that the discussions at the meeting, or subsequently, resolved all disagreements or led to some overarching agreement about the relations of these three constructs, their antecedents, and their consequences. Indeed, as the pages in this book indicate, there remain significant and theoretically important divergences in the points of view espoused on these issues by the leading researchers in the area. Yet what was clear is that the discussion at the meeting and beyond led to major clarifications in points of view and a recognition of the theoretical issues along which viewpoints diverged. Not all was divergence of opinion, however. For instance, there was no doubt that all at the meeting agreed about the excellence of Trappist breweries and of Belgian cuisine. Additionally, and importantly for the present volume, we all agreed that continuing dialogue on these important issues, without or with the assistance of Belgian ales and cuisine, would be beneficial. After returning to our respective institutions and laboratories, a large number of participants contacted us as the meeting’s organizers and suggested that further dialogue might be useful in terms of a collection of papers that we might publish jointly. This would not only facilitate further communication among the people at the meeting, but it might also expose a
xiii
broader audience to the issues that we had found so compelling. In the months that followed, we asked participants at the meeting to contribute to the present edited volume, and we also approached a number of people whose work was clearly relevant but who, for various reasons, were unable to attend the meeting. The end product is the present edited volume which brings together 22 chapters authored by no fewer than 62 researchers coming from the leading research centers in North America, Europe, and Australia. The book would not have been possible if it were not for the help and support of numerous people. We would like to take this opportunity to thank the graduate students and staff members who contributed to making the small group meeting a memorable event both scientifically and socially. In particular, we would like to thank the undergraduate and graduate students of the social psychology division of the Université Catholique de Louvain at Louvain-la-Neuve, David Bourguignon, Michel Desert, Stephanie Demoulin, Muriel Dumont, Claudia Estrada, Steve Rocher, as well as our division secretary, Anne-Françoise Cabiaux. We also want to acknowledge the generous sponsorship of the European Association of Experimental Social Psychology, the Belgian Fonds National de la Recherche Scientifique, and the Department of Psychology of the Catholic University of Louvain at Louvain-la-Neuve. We are also indebted to Allison Mudditt, then at Psychology Press and Stacy Mayill for their enthusiastic response to the project. As publisher, Psychology Press offered its full backing at all stages of the process. And, as editors, we also want to thank all of the authors of the chapters in this volume for having the wisdon to give us their distinguished contributions and the patience to bear with us as this project took unreasonable amounts of time. Last but not least, we would like to thank our families and, especially, our partners in life, Isabelle, Liz, and Florence for their constant support.
1 Perceived Variability, Entitativity, and Essentialism Introduction and Overview VINCENT YZERBYT Catholic University of Louvain at Louvain-la-Neuve CHARLES M.JUDD University of Colorado OLIVIER CORNEILLE* Catholic University of Louvain at Louvain-la-Neuve *Also at the Belgian National Fund for Scientific Research
This his volume represents an attempt to provide an integration of three different lines of research in social psychology. These three are reflected in our choice of title: “Perceived Variability, Entitativity, and Essentialism.” Each of these three terms emerged from distinct lines of work, much of it done by contributors to this volume. As we will see, these three terms, when applied to the perception of social groups, refer to related concepts. Our goal in launching this volume was to clarify their definitions, their relations, their causes, and their consequences. One line of work, the Entitativity part of our title, derives from a seminal theoretical article that Donald T.Campbell published in 1958 on the perceptual reality of social groups. He suggested that groups may vary in the extent to which they are perceived as being a “real thing” or an entity, and he coined the rather unfortunate term entitativity to refer to this characteristic. He went on to speculate about cues that perceivers might use to infer the entitativity of groups. Such cues were thought to include the extent to which group members are similar to each other, whether group members function in a coordinated manner in pursuit of shared goals, whether group membership is stable and group boundaries relatively fixed, and whether group members are physically near each other. Although Campbell offered no data in support of his ideas about these perceptual cues, a number of subsequent researchers have carried his work on group entitativity forward (e.g., Abelson, Dasgupta, Park, & Banaji, 1998; Brewer & Harasty, 1996; Dasgupta, Banaji, & Abelson, 1999; Hamilton, Sherman, & Lickel, 1998; Lickel, Hamilton, Wieczorkowska, Lewis, Sherman, & Uhles, 2000; S.J.Sherman, Hamilton, & Lewis, 1999). This work has clearly demonstrated that indeed groups do differ in the extent to which they are perceived to be entities and that many of the perceptual cues that Campbell outlined are in fact attended to in forming such perceptions. Independent of this work on groups as entities, stereotyping researchers have been interested in the strength of group stereotypes, focusing in particular on perceptions of group variability, the second term in our title.
2
THE PSYCHOLOGY OF GROUP PERCEPTION
Presumably, if perceivers have particularly strong stereotypes about given social categories, then they should see relatively little diversity among the members of such categories. The attributes associated with the category as a whole should also be associated with nearly every category member, allowing strong inferences to be made from the group to individual members and vice versa. Because of the fact that stereotypes have traditionally been associated with significant social categories (e.g., gender or ethnicity), nearly all of this work has focused on the perceived variability of such categories rather than other sorts of social groups (e.g., task groups, intimacy groups). Additionally, most of the work on perceived variability has focused on one particular determinant of perceived group variability, namely whether the target group is an ingroup or an outgroup. A robust, although not universal finding is the outgroup homogeneity effect, whereby perceivers judge ingroups to be more variable than outgroups (e.g., Judd & Park, 1988; Linville, Fischer, & Salovey, 1989; Park & Rothbart, 1982; Park & Judd, 1990; Quattrone & Jones, 1980). More recently, research has focused on the ways in which power, status, and category size (minority versus majority) may moderate these ingroup/outgroup differences in perceived variability (e.g., Simon & Brown, 1988; Lorenzi-Cioldi, 1988; Guinote, Judd, & Brauer, 2002). The third line of work that we seek to integrate derives from work in cognitive and developmental psychology on categorization processes. This work has differentiated between two different kinds of categories, typically labeled natural kinds and human artifacts. The former are categories where there is an essence, often biological in nature, that defines and dictates category membership. The latter consists of categories that are constructed by perceivers in the absence of any inherent or biological basis for categorization. The prototypic natural kind category is a species (e.g., leopards, elephants). Human artifact categories are exemplified by object categories, such as chairs or tables. Cognitive and developmental psychologists have argued that perceivers routinely dis tinguish between these two different sorts of categories and reason about them in rather different ways (e.g., Gelman, 1988; Hirschfeld, 1996; Keil, 1989). Based on this work, social psychologists have recently become interested in the extent to which social groups and categories are “essentialized” (the third term in our title) or seen as natural kinds. For instance, Rothbart and Taylor (1992) and Yzerbyt and colleagues (Yzerbyt, Corneille, & Estrada, 2001; Yzerbyt, Rocher, & Shadron, 1997; Yzerbyt & Rogier, 2001) have argued that perceivers often treat social categories as natural kinds, assuming there exists some underlying essence that unites category members. Haslam and colleagues (N.Haslam, 2002; N.Haslam, Rothschild, & Ernst, 2000) have explored the components and consequences of essentialized views of groups. And Levy, Plaks, and colleagues have examined individual differences in the tendency to see social categories as natural kinds (Levy, Plaks, Hong, Chiu, & Dweck, 2002; Levy, Stroessner, & Dweck, 1998; Plaks, Stroessner, Dweck, & Sherman, 2001). A few years ago, when we decided to initiate this volume, it seemed evident to us that the central constructs from these three research traditions were related. Groups that were perceived as entities probably also were groups where the perception of variability was low and were also probably more likely to be essentialized. Yet the literature to integrate theoretically these three was remarkably absent. Our hope is that this volume now fills this void. The chapters in this volume are grouped into two sets. Those in the first we consider to be more definitional and theoretical in nature. Here authors who have been active contributors to one or more of the above traditions have attempted to theoretically delineate these constructs and their interrelations. The chapters in the second set are a bit more empirically oriented, and they tend to explore the causes and consequences of entitativity, perceived variability, and essentialism in the perception of groups. In the sections that follow in this chapter, we briefly provide a road map for chapters in both sections. In so doing,
PERCEIVED VARIABILITY, ENTITATIVITY, AND ESSENTIALISM
3
we also hope to provide our own perspective on the theoretical integration of these three terms and their research traditions. SECTION ONE: DEFINITIONAL CONSIDERATIONS We see the chapters in this section as struggling with the issues of (a) what exactly do we mean by the terms perceived variability, entitativity, and essentialism, and (b) how are these related to each other, both theoretically and empirically. As will be clear on a close reading of these chapters, our authors have rather different perspectives on these two issues and argue for theoretical integration in rather different ways. Although this diversity ultimately threatens our goal of providing a quick and easy integration of these traditions and literatures, it does mean that the reader must attempt his or her own unique theoretical understanding, and we see this as a very positive outcome indeed. The initial chapter in this section, by Marilynn Brewer, Ying-yi Hong, and Qiong Li, focuses in particular on one of the central attributes that D.T.Campbell (1958) recognized as a cue in the perception of entitativity: whether or not a group has a set of shared goals and purposes and acts in a coordinated manner toward those. Brewer et al. argue that the perception of similarity or homogeneity of group members is not a sufficient condition for the perception of group entitativity. Rather, entitativity is inferred either when the group is a seen as a natural kind (essentialized) or when it is seen as acting in a coordinated manner on common goals and purposes. And these two bases for the perception of entitativity have somewhat different consequences, with the perception of essence influencing more judgments about individual group members and the group prototype, and the perception of agency or common goals influencing emotional reactions to the group (and its purposes) as a whole. Beyond these theoretical considerations, Brewer and colleagues report some data from cross-national samples in which perceptions of both a group essence and group agency are shown to predict judgments of entitativity. Although they report some cultural differences, the bottom line is that both are influential in both cultures. Like Brewer et al., David Hamilton, Steven J.Sherman, and Julie S.Rodgers primarily focus on the perception of group entitativity and how it relates to the other constructs of essentialism and perceived homogeneity. They argue that these three concepts, while they are certainly positively correlated, are not redundant. Further they argue that the exact relations among them depend on the type and function of the group involved. Specifically, their empirical work has differentiated among social categories, task groups, intimacy groups, and loose associations (e.g., people waiting for a bus together), and the relative degree of entitativity, homogeneity, and essentiality for all four types of groups. They suggest that intimacy groups (e.g., families) are the highest on entitativity, but typically do not have strong stereotypes associated with them. They suggest that social categories may be essentialized, but they are actually fairly low on perceived variability compared to the other types of groups. For task groups, they argue that perceptions of entitativity are relatively high, but again essentialism and perceived homogeneity may be relatively lower than for other types of groups. The bottom line in the Hamilton et al. chapter is that while these three constructs are related, they are clearly discriminable, with different causes and consequences. By focusing on the four different types of groups that their research has uncovered, they characterize the differences between entitativity, perceived homogeneity, and essentialism primarily by discussing how they vary (and covary) between types of groups, rather than how they vary (and covary) across groups within a given type. This is a somewhat different approach than that followed by the authors of the other chapters in this section.
4
THE PSYCHOLOGY OF GROUP PERCEPTION
The Brewer et al. and Hamilton et al. chapters seem most focused on entitativity and its relations to other constructs. Nick Haslam, Louis Rothschild, and Donald Ernst, on the other hand, clearly come out of the research tradition on essentialism and the properties that differentiate essentialized groups from groups that are less likely to be seen as “natural kinds.” They start their chapter with a fascinating tour of the history of thinking about essentialism, tracing the origin of the concept in the linguistic and philosophical traditions. They then report on a series of empirical studies that have explored the characteristics of essentialized groups. From this work, they argue that there are two components to essentialism, with groups varying more or less independently on them both. Characteristics included in the first component have to do with traditional notions of what constitutes a “natural kind”: they are groups that are immutable, natural, historically invariant, have sharp boundaries, and necessary features. The second component, they argue, is made up of characteristics traditionally associated with entitativity. Characteristics such as homogeneity, inductive potential (i.e., inferring what someone is like from group membership and vice versa), and membership based on shared inherent properties are associated with the second component. Starting from a theoretical interest in defining the characteristics of essentialized groups, Haslam, Rothschild, and Ernst thus come to the conclusion that there are two components: one having to do with groups being seen as “natural kinds” and the other having to do with the perceived entitativity of groups. Thus, they see entitativity as one component of essentialism, and, within entitativity, perceived homogeneity is one factor that leads to the perception of entitativity. This is clearly, then, a sort of nested structure, with perceived homogeneity as a condition for entititativity and entitativity as a condition for essentialism. While Brewer et al. and Hamilton et al. start from the point of view of explaining entitativity, and Haslam et al. start from the point of view of explaining essentialism, Mick Rothbart and Bernadette Park emerge much more squarely out of the stereotyping literature. Accordingly, their initial question is focused on the process by which stereotypes may be changed, and in this regard they necessarily focus on factors that affect whether individual group members who disconfirm a group stereotype will be functionally integrated into the group and thereby lead to change. They suggest that whether a disconfirming group members is functionally included in or excluded from the group depends on both the perceived variability of the group and the degree to which the group is seen as an entity. They further argue that entitativity is not simply the same thing as perceived variability. Group entitativity is also affected by the extent to which a group is seen as acting on common goals or its degree of agency (consistent with Brewer et al.). Finally, they suggest that essentialized groups are a subset of groups that are seen as entitative, having both high perceived similarity and a sense of agency. All of this leads them to suggest that disconfirming group members will be functionally excluded, and stereotype change inhibited, when group boundaries are strong. And this is most likely to be the case in entitative groups that possess shared attributes and common goals, and are essentialized. The final chapter in this section, by Vincent Yzerbyt, Claudia Estrada, Olivier Corneille, Eléonore Seron, and Stéphanie Demoulin, also starts by asserting that these three constructs are not one and the same thing. But they argue in particular for a different relationship between entitativity and essentialism than that articulated by others. They suggest that entitativity is the observable or phenotypic attribute of “groupness,” whereas group essence is the underlying or genotypic attribute that results in the phenotypic expression. They then show, in a series of reported studies, that when essentialism is manipulated, entitativity and perceived similarity are consequences. Thus, perhaps like Rothbart and Park, they suggest that essentialism is the most fundamental of the three concepts, that essentialism requires perceived similarity and entitativity, and that perceived similarity and entitativity are broader constructs and, as such, can be found with nonessentialized groups.
PERCEIVED VARIABILITY, ENTITATIVITY, AND ESSENTIALISM
5
This brief summary of these chapters (admittedly not the summary that these authors themselves might have provided) clearly points to diversity in both the author’s starting points and the conclusions they eventually reach. Yet we believe that there are threads of consensus that are apparent across the chapters, and hopefully they show the way to a more integrative account of the theoretical relationships among entitativity, perceived variability, and essentialism. First, no one argues that these are one and the same construct. Groups vary in the extent to which they are perceived to be real real or meaningful, the extent to which their members are seen to be diverse, and the extent to which they are seen to have an essence that defines them as a “natural kind.” Although these perceptions certainly covary, they are not one and the same thing. A necessary corollary of this is that these three characteristics of groups have at least somewhat different causes and consequences. The factors that cause a group to be seen as entitative, that lead to perceptions of homogeneity, and that result in an essentialized view of group membership are not one and the same. And yet we would suggest that all three of these perceptions have the potential for influencing each other. Many of the studies that are summarized in these chapters suggest that each of these perceptions has potential implications for the others. Manipulations of essence lead to higher perceived similarity and higher entitativity. Groups that come to be seen as less variable are more likely to be seen as entitative, and so forth. Though not identical constructs, and though potentially having different causes and consequences, they nevertheless are causally related to each other. And the causal relations are certainly reciprocal and nonrecursive. If we grant this, then the next obvious question concerns whether these constructs share a nested or hierarchical structure of relations or whether they are simply overlapping sets, capable of influencing each other in a probabilistic manner. Here, we are taken by many of the ideas contained in Figure 1 of Rothbart and Park. Consistent with Brewer et al., it seems that entitative groups derive their “realness” either from the perceptions of similarity of members or from a sense of agency and purpose. And the conjunction of the two is particu larly likely to lead to group entitativity. But essentialism seems to us to be more than just another partially overlapping set of groups. Groups that are essentialized seem to us to necessarily have a realness to them that cannot be denied, and that realness inevitably leads to strong and impermeable group boundaries with high perceived genotypic similarity. In this sense, essentialism seems to have a rather different conceptual basis from the other two constructs, in that it implies attributional consequences that the other two do not. Perceived homogeneity focuses on the extent to which group members share attributes. Entitativity focuses on the extent to which a group is seen as real or coherent. Essentialism derives from both of these but it seems to us to imply more. It entails not just a description of a group and what makes it a group, but also a theory about why group membership matters and what it implies. If a group is essentialized, then we are permitted to make inferences about why this group exists (why it is an entity) and why group members are similar to each other. We are not saying simply that the group is an entity and that its members are similar, but we are implying an attributional process for the origins of entitativiy and similarity: this is a group and its members are similar to each other because they all share an essence that is invariant and unmutable. Essentiality, in other words is not just the conjunction of similarity and entitativity. It really implies a theory about the origins of similarity and entitativity. It implies that the perceiver has a “natural kind” explanation for the existence of the group and for why its members are similar to each other. It suggests that the perceiver attributes groupness and similarity to internal, dispositional, and immutable characteristics of group members. Turning to the distinction made by Hamilton et al. among different types of groups, it seems to us particularly interesting that it is often social categories (rather than task groups or intimacy groups) that are imbued with essentialistic qualities. And then the interesting question for us concerns not the differences
6
THE PSYCHOLOGY OF GROUP PERCEPTION
among these types of groups (categories versus intimacy groups versus task groups) but the conditions under which social categories are essentialized. This focus on social categories, and the question of when they are essentialized, enables us to effectively bridge the three diverse traditions that we identified at the start of this preface and that gave rise to our notions of entitativity, perceived variability, and essentialism. Strong stereotypes, prejudice, discrimination, and hostile intergroup relations have been issues at the core of social psychology since its founding. And from our point of view, notions of entitativity, perceived variability, and essentialism take on particular interest to the extent that they help us understand these social problems. Strong stereotypes occur when the members of social categories are viewed as very similar to one another and when the boundaries that differentiate those categories are seen as sharp and fixed. And then when perceivers go the additional step of essentializing those categories, then the definition of a social stereotype takes on a very different aspect: category membership becomes inevitable and immutable. Category members resemble each other because of their unchangeable and true essence. And the boundaries that divide groups and that lead to social conflict are seen as fixed forever as a natural part of the world, just like the lines that separate different species. Such essentialized social categories, it seems to us, are what leads to the most severe forms of intergroup conflict and hostility, and ethnic cleansing, or ethnic genocide, when it has occurred historically, has certainly been accompanied by an essentialistic rhetoric. From this point of view, what becomes important is the evolution of essentialistic thinking about social categories: the factors that lead to more essentialistic views of social categories, given perceived entitativity and similarity. And this leads us to want to understand the dynamic relations among these constructs and the factors that lead to changes in all three, within the same social categories, over time and across perceivers. The chapters in the first section of the book, discussed above, really focus more on defining the three constructs and to a lesser extent on their dynamic interrelations. The chapters in the second set, which we now briefly summarize, are concerned more with the various factors that lead to the perception of group entitativity, group homogeneity, and essentialism. They also stress some of the consequences of entitativity and essentialism. And from our point of view, when talking about social categories, these are the important questions to be addressed in the attempt to understand why it is that groups and people conflict and how such conflicts might be ameliorated. SECTION TWO: VARIATIONS ON THE THEME As we have argued above, the perspective prevailing in some of the definitional chapters is that the perception of homogeneity, entitativity, and essentiality depends largely on the particular group that is being appraised (Hamilton et al., this volume; Haslam, this volume; but see Yzerbyt et al., this volume). This target-based approach has proven fruitful in stressing some major differences among groups. At a very general level, two images of social groups seem to emerge. Whereas some groups elicit the belief that their members are closely connected, well organized, and working together to reach some joint outcome, others are associated with the idea that similarity of group members is the key feature, some deep underlying marker making it difficult if not impossible to change or deny group membership. From this point of view, a perceiver’s impression of a group depends on the particular characteristic of that group. With a few exceptions, the chapters in this second section tend to stress a rather different approach, one that we see as most promising. Rather than assuming that groups have defining characteristics that affect how they are perceived, most authors of the chapters in this second section concentrate on possible structural and social factors that may shape a perceiver’s propensity to see any given group as a coherent entity or as a looser set of people. A neat illustration of this perspective is the chapter by Jason E.Plaks, Sheri R.Levy, Carol S.Dweck, and Steve J.Stroessner. These authors build a strong case for an individual
PERCEIVED VARIABILITY, ENTITATIVITY, AND ESSENTIALISM
7
difference approach that complements the view according to which entitativity or perceived variability primarily derives from the intrinsic features of the target groups. To be sure, most if not all social psychologist would agree that the stimulus is not the sole factor in the perception equation and that the perceiver plays a most critical role. Still, according to Plaks and colleagues, most contemporary efforts may have somewhat underestimated the extent to which group perception also resides in the eye of the beholder. People’s a priori beliefs about human nature, whether chronic or more transitory, should likely orient their perception of entitativity and homogeneity. In line with this idea, Plaks and colleagues identify two contrasting lay theories that underlie quite divergent perspectives for understanding individuals as well as groups. Whereas “entiteists” view personal characteristics as fixed entities that are unlikely to change over time and they expect a high degree of consistency in people’s behaviors, “incrementalists” prefer to see personal characteristics as dynamic and open to modification, and they situational forces that impinge on people’s actions. Plaks and colleagues provide an impressive body of evidence to suggest that these two views translate into distinct takes on group perception. Studies indeed reveal that, compared to “incrementalists,” “entiteists” more readily apply psychological traits to groups, perceive less within-group variability and greater between-groups differences, form and endorse stronger stereotypes, and process information about the group members in such a way that less attention is given to the counterstereotypic evidence. Entity theorists not only use and preserve their stereotypic beliefs more so than “incrementalists,” they also are more likely to believe that innate factors rather than shared environmental experience are accountable for shaping group members’ characteristics (see also C.L.Martin & Parker, 1995; Yzerbyt & Rogier, 2001). In contrast, Plaks and colleagues argue, “incrementalists” are more sensitive to the presence of shared goals as a cue to group coherence. There are a number of links that can be drawn between the contribution of Plaks and colleagues and other chapters in this volume. In particular, we would point to the overlap between their analysis and Brewer’s distinction between groups organized around some essence and those organized around some goals. Also, we see interesting parallels between Plaks et al.’s suggestions and those made by Social Identityy Theorists (SIT) about stability and permeability, entity theorists living in a world where groups are clearly segregated and the social structure is seldom challenged and “incrementalists” believing the social landscape is largely open to revision. In addition to perceiver differences, it is clear that a number of other factors affect the perception of group homogeneity, entitativity, and, possibly, essence. Most of the remaining chapters help us identify these factors. To be sure, these factors can be grouped in various ways. One way that we find rather convenient in organizing the various chapters builds on the extent to which the perceiver plays an active role in the construction of entitativity. At one end of the spectrum, one can identify a set of factors to which the perceiver simply responds. He or she plays no real active role or does not engage in what could be seen as truly motivated or strategic group perception and cognition. Familiarity with a group is a good illustration of this kind of factor: Differences in group familiarity may affect the perception of group entitativity and homogeneity, but these differences are not actively sought out or strategically motivated on the part of the perceiver; rather he or she just happens to have more or less knowledge about the group, and this affects perceptions. Going to the opposite end of the continuum, the perception of entitativity might be seen to result from a deliberate line of actions and cognitive strategies. For instance, if the perceiver is motivated to think of his or her own group as needing to realize some personally important goal, to reach a specific end, then actions may be undertaken to organize the group toward that purpose, and more entitative group impressions will
8
THE PSYCHOLOGY OF GROUP PERCEPTION
ensue. At this end of the continuum, then, the perceiver actively and strategically affects the group itself in ways that lead to changes in how entitative it is seen to be. Halfway along the continuum, one might find that processes purposively engaged in by the perceiver might end up in promoting the perceived entitativity of the groups even though this was not the perceiver’s primary goal. For instance, sometimes people talk with each other about their own groups and other groups. The goal of communication is not to strategically affect the perceived entitativity of those groups. But it may still have these unanticipated outcomes. In the remainder of this introductory chapter, we would like to walk you through this continuum from one pole to the other, discussing each of the chapters in turn. Needless to say, this tour only provides a highly subjective and necessarily partial account of the content of the various chapters. Having reviewed all the chapters, we will then come back to a number of issues that allow us to compare and to contrast more specific contributions. A first factor that may affect the extent to which perceivers see groups as being coherent wholes is the very act of categorization. By slicing the social environment into different zones of coherence, perceivers may well end up overestimating the cohesiveness of the groups they have categorized. In short, the mere act of categorization is likely to promote similarity among the members of same category and, as a consequence, entitativity of the category. This observation is of course far from new, and the single most famous empirical demonstration of the phenomenon dates back to the study conducted by Tajfel and Wilkes (1963). The original paradigm has come in many disguises. The most popular version is the so-called “Who said What?” paradigm in which participants observe a discussion between a limited number of members comprising two distinct social categories, black versus whites, males versus females, overestimators versus underestimators, etc. At the end of the discussion comes a surprise memory test in which observers are presented with individual statements and asked to assign them to the correct speaker. Categorization is assumed to have affected the perceptions of homogeneity if errors of recognition more frequently involve within-category confusions than between-category confusions. In their chapter, Karl Christoph Klauer, Katja Ehrenberg, and Ingo Wegener take issue with this working assumption and provide a thorough examination of the various components that may be responsible for memory errors in experimental settings that rely on the “Who said What?” paradigm. Instead of comparing the number of between-categories and within-category confusions, the authors suggest that the paradigm be altered by adding a set of distracters to the list of statements comprising the memory test. Doing this would allow researchers to disentangle the impact of a number of independent cognitive processes, some directly relevant to the issue of homogeneity and some not, and to examine the impact of various contextual factors thought to affect the emergence of homogeneity. The relevant cognitive processes are person memory (Is there a memory of the person who made the statement?), category memory (Is there a memory of the category of the person who made the statement?), and category guessing (Are there cues that can be used to infer the category membership of the person making the statement even though category membership cannot be recalled?). Whereas accurate person memory within a category is indicative of heterogeneous processing of the discussants, both category memory and category guessing suggest rather different processes indicative of homogeneity in perception. Klauer and colleagues present a series of studies aimed at validating the relations between the process indicators and the underlying mental operations. For instance, a manipulation of the salience of the categories only affected the category memory parameter and not the other indices. Beyond establishing the convergent and discriminant validity of the various process measures, the authors examine the impact of a series of important contextual factors on group perception such as relative group size and crossed categorization. Their findings lead them to argue for the dissociation between various facets of category-
PERCEIVED VARIABILITY, ENTITATIVITY, AND ESSENTIALISM
9
based processing, possibly allowing a reconciliation of contradictory findings with regard to the impact of cognitive load in social perception. Keeping with the discussion of some fundamental processes involved in the categorization processes, Craig McGarty considers how prior expectations about the existence of group differences interacts with stimulus information one receives about groups to affect the extent to which those groups are seen as entitative. He argues that the mere existence of prior beliefs that says that groups differ from each other is sufficient to trigger the formation of group stereotypes and the perception of groups as entities, assuming that in fact there is sufficient information received about those groups to enable effective categorization. McGarty uses these general ideas to offer an insightful explanation for illusory correlation results, suggesting some nice extensions to Fiedlers (1991) and E.R.Smiths (1991) account based on information loss and unreliability of judgment. McGarty’s explanation suggests that the typical illusory correlation results are due to the expectation that the groups must differ in some way and then by the availability of sufficient information about the two groups to permit that expectation to result in rather different stereotypic beliefs. In a similar vein, McGarty suggests that perceivers may hold prior expectations about the degree to which social categories are essentialized. And, consistent again with his theoretical position about the constraints that exist between expectations and the available stimulus information, he suggests that these expectations may permit essentialized beliefs for some social categories more than for others. Continuing with the idea that the perceived entitativity of groups may emerge simply from extraneous factors that involve little strategic or motivated reasoning on the part of the perceiver, Patricia W.Linville and Gregory W. Fischer extensively discuss the role of differential familiarity with groups as a factor that affects their perception. Interestingly, rather than examining a variety of prototypic attributes thought to characterize groups in general, these authors concentrate on a somewhat neglected facet of the mental representation of groups, namely the perception of covariation among the features describing category members (but see Judd & Lusk, 1984; Park, Ryan, & Judd, 1992). This is an important consideration because, as Linville and Fischer argue, perceived covariation is likely to be related to the degree to which people make distinctions among group members. Moreover, there is no reason to expect any strong link between the perceived covariation among a set of attributes and the perceived variance along each individual attribute. As much as we have learned from closer scrutiny of the perception of single feature variability, it may be fruitful to examine the perception of multiattribute covariability. Linville and Fischer’s contribution revolves around two main ideas. First, perceptions of covariation generally overestimate the actual level of covariation, and this tendency is stronger at lower levels of familiarity. Supportive evidence for this conclusion comes from several studies using various broad social categories such as age, gender, occupation, or even race groups. When people have extensive contact with members of the target group, when they are members of the group, or when they are otherwise knowledgeable about the group, their judgments entail lower perceived covariation among the various attributes. Additional work on highly entitative groups such as basketball teams confirms this pattern and additionally indicates that familiarity and expertise allow people to make less biased estimates of objective feature covariation. Interestingly, Linville and Fischer’s strategy for collecting information about perceived covariation is highly implicit in that participants are not directly asked to make covariation judgments. Rather, perceived covariation is computed from judgments that are given along trait dimensions. Second, Linville and Fisher suggest that implicit perceptions of covariation play a role in social judgment inasmuch as a higher degree of covariation should be associated with more extremity in judgments. Again, data from high and low entitative groups seem to be consistent with this conjecture. Linville and Fischer conclude by saying that perceived covariation need not be related in any simple way to group entitativity.
10
THE PSYCHOLOGY OF GROUP PERCEPTION
Depending on the basis for entitativity, either similarity or interdependence, one might expect more or less perceived covariation. Unlike Linville and Fischer, who focus almost exclusively on familiarity, Mark Rubin, Miles Hewstone, Richard J.Crisp, Alberto Voci, and Zoë Richards additionally tackle the factors of group size and group status and examine how all three factors play a role in the emergence of the outgroup homogeneity effect, i.e., the tendency to perceive the outgroup as being more homogeneous than the ingroup. On the basis of the evidence obtained in minimal group studies and with gender groups, Rubin and colleagues question the status of differential familiarity as a necessary cause for the outgroup homogeneity effect to emerge. In fact, and contrary to both the differential familiarity hypothesis and Lorenzi-Cioldi’s position on gender effects, gender outgroup homogeneity (OH) can be seen as a modest yet reliable phenomenon (Park & Judd, 1990). The authors then engage in a critical analysis of Lorenzi-Cioldi’s interpretation of the larger OH effect among men than among women. On the basis of the accumulated evidence, they conclude that members of high status groups, and thus men, perceive greater outgroup homogeneity effects. In contrast, members of low status groups, and women in particular, show no such tendency. Rubin and colleagues also discuss the so-called minority group homogeneity effect as well as the interactive effect of group status and group size. In the second part of their chapter, Rubin and colleagues present a research program aimed at disentangling the impact of these three factors on the perceived dispersion of members of gender groups. Their studies reveal that group size is indeed more strongly associated than status or gender with a tendency to perceive outgroup homogeneity, familiarity playing no causal role in the process. Interestingly, however, measures that make reference to individual group members (dispersion rather than global similarity) and targets that stress interpersonal comparisons (individual targets rather than abstract categories) suggest that men are more prone than women to enhance their personal identity, a tendency that facilitates the emergence of outgroup homogeneity among men. Ana Guinote’s chapter presents yet another perspective on the role of group size and group status on the perceived homogeneity of groups, adding the factor of power. The empirical work that is being reviewed rests on two main hypotheses. For one thing, Guinote argues that the position of individuals in the social structure directly affects their perception of control. This perception of control, in turn, influences the extent to which perceivers will attend to the variability of the groups out there. Being able to count on numbers, status, and power thus leads people to see other groups in a more homogeneous manner. For another, Guinote holds that perceived control also has some bearing on the actual spontaneity in behaviors, that is, people’s objective heterogeneity. Members of large, high-status, or otherwise powerful groups are likely to act in more deliberate ways and thus be inherently more variable than members of small, low-status, or powerless groups. These two phenomena converge so that the members of subordinate groups come across as less variable than the members of dominant groups. A series of studies are presented that lend support to these two hypotheses. With the two following chapters, we continue the journey along our continuum. After concentrating on homogeneity and entitativity as the accidental result of the peculiarities of people’s perceptual and cognitive processes or their mere position in the social structure, we now turn to work that assigns a more active role to social perceivers. Indeed, a number of contributions see homogeneity, entitativity, and possibly essence, as emergent products of social interaction. As a case in point, Markus Brauer, Charles M.Judd and Micha S.Thompson tackle the important question of whether communication about a group may indeed affect the perceived entitativity of a target group. Building upon Park and Hastie’s (1987) distinction between instance-based versus abstraction-based stereotype acquisition, the first study uses a rumor transmission paradigm to show that people end up holding more extreme and less variable views about a target group when they learn about the group only
PERCEIVED VARIABILITY, ENTITATIVITY, AND ESSENTIALISM
11
from other people than when they form their impression solely on the basis of first-hand behavioral observations. Brauer and colleagues further show that this communication effect is not restricted to recipients of communication but that people asked to tell others about their impressions of a group similarly form more stereotypic views than people simply asked to study the materials. Relying on actual group communication settings, two additional studies confirmed the fact that these two effects combine to polarize stereotypes about a target group but only when all group members possess an equal (small) share of counterstereotypic information. Clearly, this chapter provides convincing evidence regarding the role of communication in the formation and preservation of stereotypes and, as such, stresses the social nature of stereotypic beliefs (for a related set of issues, see McGarty, Yzerbyt & Spears, 2002). In the following chapter, Yoshihisa Kashima further stresses the idea that beliefs about groups are social constructions. Indeed, this contribution focuses on the psychological processes involved in the production and reproduction of social reality. A central tenet of Kashima’s analysis is that communication contributes to providing a sense of realness to whatever is the topic of the exchange. Importantly, communication takes place between people in the context of a larger cultural background. In line with the theme of the present book, this phenomenon applies to social categories as much as to anything else. That is, within a given cultural context, communication makes social categories become entitative and real. As a consequence, Kashima argues, people tend to endow social categories with some essence. A noteworthy feature of Kashima’s argument is the distinction between the notion of entitativity as proposed by D.T. Campbell, which rests on a series of “measurable” perceptual clues and refers to the ontology of groups, and the concept of essentialism, which more directly relates to the psychological sense of realness and is highly sensitive to the occurrence of communication (see also Yzerbyt, Corneille, & Estrada, 2001). Several pieces of empirical evidence allow Kashima to argue that communication indeed not only produces more polarized and homogeneous impressions of groups but also leads perceivers to see the characteristics of the target group to be more unalterable. Additional cross-cultural work also shows the nonnegligible role of culture in the attribution of essence to social entities. Kashima’s chapter is particularly important in that it simultaneously stresses the active part people play in constructing social reality and the critical role of communication in providing the psychological sense of that reality. The remaining chapters in the book carry us to the more strategic end of the continuum. For many of the authors of this third category of chapters, entitativity comes across as a feature that is being monitored, relied upon, used, and even sought out and constructed. To be sure, we note a progression in the extent to which the authors subscribe to such a functional perspective in their treatment of group entitativity. Starting with the idea of a genuine interest in the perception of regularities among the group members, one ends up with some indication that people are indeed doing their best to construct the perception that a group is a coherent social entity. Carey S.Ryan provides a good demonstration of people’s special attention to aspects of group homogeneity and entitativity. Her chapter focuses on the process by which individuals become socialized into groups, specifically concentrating on the ways in which socialization affects the perception of group variability, particularly of the ingroup. According to Ryan et al., new group members initially prefer to pay attention to the similarities of individuals within the ingroup at the expense of their idiosyncrasies. Why is this? According to Ryan, discovering what the common themes are among ingroup members allows new members to deal with their uncertainty and lack of knowledge about the group and maximizes their chances of acceptance. At the intergroup level, such a group-level approach helps them to map the social landscape and to clarify the distinctive features of their ingroup. At the intragroup level, the construction of a schematic understanding of the group provides a means to reassure the other members that one is a good recruit. Ryan’s model is thus essentially predicting a stronger need for a homogeneous view of the ingroup
12
THE PSYCHOLOGY OF GROUP PERCEPTION
at early stages of group membership. As time passes, a greater increase in perceived dispersion is found for ingroups than for outgroups, meaning that outgroup homogeneity should emerge over time. Data from a longitudinal study with new members of sororities lend strong support to these conjectures. Ryan et al. also presents experimental evidence confirming that people who join new groups and are uncertain about their entry in the group primarily search for information about the group rather than about the individual members of the group. As a whole, the data presented by Ryan et al. are unique in examining dynamic impressions of groups over time. They show that forming an individualized view of group members may not be the most functional way to gain access to and become a good member of a group. Instead, constructing some global understanding of one’s group by spotting the similarities is an efficient strategy, at least in the initial phases of membership. Only with the passage of time does the focus on the formation of an adequate representation of the group as a whole give way to some concern for individual differences within the group. Russell Spears and colleagues also focus on homogeneity as a valued characteristic of the group. Adopting a social identity and self-categorization framework, these authors focus on the concept of distinctiveness rather than entitativity. According to them, distinctiveness may well entail a stronger emphasis on the motivational role of the social self and indeed seems to be more directly tied to the consideration of an intergroup context, two features which also have a number of consequences at the measurement level. Having said this, Spears and colleagues set out to examine the different functions of group distinctiveness. In a nutshell, distinctiveness is thought to provide group members with a sense of identity (“who are we?”) as well as with a way of dealing with unfavorable social status (“what can we do?”). As far as the identity function is concerned, Spears and colleagues make a number of suggestions that are strongly reminiscent of other views presented in this volume. Specifically, they note that the perception of the distinctiveness of a group can be as much a reflection on the nature of things as a reaction to the lack of structure and clarity in the social environment. Indeed, empirical evidence is presented showing that group members may sometimes be motivated to differentiate their group from other groups. As to factors that promote such a response, Spears and colleagues suggest that the type of groups and the visibility of other group members play a key role here. Turning to the instrumental function, Spears and colleagues propose that distinctiveness is likely to be useful in the pursuit of group goals that require coordinated group action. Not surprisingly, the presence of a group project should be critical in promoting the willingness of group members to emphasize the coherence and cohesiveness of the group. Interestingly, the authors note that these two functions of distinctiveness are complementary: Distinctiveness/entitativity as ends is a prerequisite for distinctiveness/entitativity as means. Spears and colleagues then go on to present an impressive series of studies that focus on the way people’s level of identification with the group moderates the search for distinctiveness, be it for identity or instrumental reasons. The idea that group entitativity may serve a group in promoting collective action is also very much present in Katherine J.Reynolds, Penelope J.Oakes, Alex Haslam, John C.Turner, and Michelle K.Ryan’s contribution. Still, these authors propose quite a different take on a number of themes that have been tackled in previous chapters. First, like Plaks and colleagues and a few other chapters in this collection, they emphasize that the emergence of entitativity is both a function of characteristics of the target groups and of the perceivers. Concerning the observed groups, the notion of comparative fit is used here to refer to the regularities observed in the social environment. Turning to the perceivers, Reynolds and colleagues stress the role of normative fit and perceiver readiness in shaping people’s tendency to appraise the social situation in terms of intergroup rather than interpersonal relations. Clearly, the notion of normative fit shares a number of features with that of naive theories. As for perceiver readiness or accessibility, one can
PERCEIVED VARIABILITY, ENTITATIVITY, AND ESSENTIALISM
13
distinguish more chronic preoccupations linked to social identification and more transient concerns deriving from the context. Reynolds and colleagues then go on to propose that structural factors such as the level of power that people enjoy and the nature of the relationship between the groups, i.e., the security of the groups status and the permeability of the group boundaries, are likely to play a major role in the emergence of entitativity. Taking issue with the findings showing that low status groups should be the ones seeing high status groups as heterogeneous, Reynolds and colleagues provide intriguing evidence indicating that low status groups confronted with impermeable and secure relations more readily engage in stereotyping as they derogate the outgroup and see it as a coherent whole to a greater extent than the members of the high status group do. As in the chapters of Guinote and Rubin et al., the work by Reynolds and colleagues makes clear that the perceived homogeneity of a group is highly dependent on the larger social context and that structural variables play a key role in triggering dynamic perceptions of entitativity. In their own way, these last two chapters stress the importance of entitativity in intergroup relations. When observers perceive a group as a coherent whole, they are likely to engage in different behaviors than when they perceive the group as a much less consistent entity. What happens when perceivers draw the conclusion that they are facing a real group rather than a collection of individuals is precisely the question addressed by Wildschut, Insko, and Pinter. Summarizing an impressive program of research, these authors build a case for the idea that groups come across as more threatening than individuals. The repeated observation of the so-called discontinuity effect in mixed-motive situations, i.e., groups are more competitive than individuals, led these authors to investigate in more detail both the content of the outgroup schema as well as the factors that contribute to its activation. As for content, Wildschut and colleagues provide convincing evidence that there is a negative schema of the outgroup consisting of beliefs or expectations that intergroup interactions are competitive, unfriendly, deceitful, and aggressive. So, for instance, communication about future moves in a mixed-motive game is less credible and persuasive when it emanates from groups rather than individuals. People also seem to believe, both at an explicit and at an implicit level, that intergroup interactions are less agreeable and more abrasive than interindividual interactions. They expect interactions with groups to be more competitive than with individuals, and, indeed, discussions involving groups include more statements about distrust than discussions between individuals. Finally, people not only experience interactions with groups as being more competitive but memory comes into play and worsens the picture even more. Turning to the activation issue, Tim Wildschut and colleagues argue that the outgroup schema will be activated whenever an aggregate of individuals is seen as an entity. A key factor in triggering the image of an entitative group is that people’s behaviors and outcomes are seen to be interrelated in producing the groups response, which the authors call procedural interdependence. Such procedural interdependence is of course reminiscent of D.T.Campbell’s (1958) notion of common fate and, Wildschut and colleagues argue, may be fostered by a variety of factors. For instance, when people are confronted with a series of individuals known to rely on consensus in order to reach a group decision, these individuals are likely appraised as a social entity. As a result, fear increases and competition steps up. All in all, the message emerging from this line of work indicates that outgroup fear may be very difficult to reduce. The fact that Wildschut and colleagues paint such a somber picture of entitative outgroups should not be interpreted as a sign that entitativity is a negative feature altogether, one that groups should see as undesirable and try to avoid. As a matter of fact, intuition and empirical evidence alike confirm that entitative outgroups are frequently seen in a negative light (Abelson, Dasgupta, Park, & Banaji, 1998), but this is not the whole story. Often, seeing a group as one coherent block is exactly what perceivers would prefer if collective action is to be initiated or if discrimination and conflict is to be rationalized. It may thus be desirable and
14
THE PSYCHOLOGY OF GROUP PERCEPTION
indeed most important that a group be perceived as an entity. In fact, for quite a few contributions in the book, the take-home message goes exactly in this direction. Interestingly, the distinctive feature is that these chapters all turn their attention to the degree of perceived homogeneity of the ingroup as opposed to the perception of the outgroup or of both groups. A good illustration of this perspective can be found in the chapter by Dominic Abrams, Jose M.Marques, Georgina Randsley de Moura, Paul Hutchinson, and Nicola J.Bown. These authors provide a thorough review of their recent empirical efforts on their subjective group dynamics approach. This line of research, which is an extension of Marques, Yzerbyt, and Leyens’s (1988) earlier work on the black sheep effect, is mainly concerned with people’s reactions to ingroup deviance. The subjective group dynamics model holds that people will simultaneously strive to maximize intergroup distinctiveness and to validate ingroup norms. Whereas the former goal is associated with a high level of category processing, the latter goal implies instead a fair degree of intragroup differentiation. Because people value membership in groups that are entitative, they should feel threatened in the presence of deviant group members who imperil the coherence of the group. Abrams and colleagues present the results of several studies showing that counternormative ingroup deviants are indeed more harshly evaluated than equally counternormative outgroup members. Moreover, they provide evidence that the phenomenon is amplified in group threatening situations. Clearly, a lack of entitativity encourages group members, especially the highly identified ones, to derogate the antinorm deviant, possibly in an attempt to try and restore the perception of group coherence and a sense of subjective reality of the group and the ingroup norms. The idea that people may want to see the ingroup as entitative is taken a step further in Emanuele Castano’s chapter. A first central idea of the chapter is both simple and far-reaching in its implications. If people like to see their ingroup as being rather entitative, they may react more positively to an ingroup that displays coherence as opposed to an ingroup for which there is a lower level of groupness. That is, entitativity occupies the role of an independent variable. Along with a number of colleagues, Castano conducted a series of studies supporting the idea that group members like it when their group is entitative. Specifically, people were found to identify more with groups that were more entitative. A noteworthy feature of this program of research is the use of a wide variety of group features used to trigger entitative perception of the group in the first place. Be it through boundedness, common fate, similarity, or salience, all the features identified by Campbell succeeded in fostering an entitative representation of the ingroup. The second part of Castano’s chapter examines two possible reasons for people’s positive reaction toward an entitative ingroup. The first reason refers to what Castano calls the terrestrial value of entitativity: Groups that are more coherent are endowed with higher intentionality. In other words, there is a sense of purposiveness that entitative groups have that make them threatening if they are outgroups and attractive if they are ingroups. The second reason is the so-called celestial value. Building upon some intriguing empirical research in which terror management theory plays a central role, Castano suggests that membership in a close-knit social entity allows individuals to extend themselves through space and time and releases them from a sense of finitude about their own lives. The chapter concludes with a discussion of the relevance of this work for a series of hot issues in political psychology and international relations. In the next contribution, Michael A.Hogg addresses a number of themes that clearly mirror the issues dwelt upon by Abrams and colleagues and by Reynolds and colleagues. They also extend some of the issues addressed by Castano. The chapter examines the reasons that may lead people to join highly entitative groups, which he calls “totalist” groups. One possible way to account for the seductive power of more entitative as opposed to less entitative groups, Hogg argues, is that such groups are particularly functional in addressing people’s needs to resolve uncertainty. The point of departure of Hogg’s line of reasoning is that,
PERCEIVED VARIABILITY, ENTITATIVITY, AND ESSENTIALISM
15
contrary to what social identity theory would want us to believe, self-enhancement is likely not the sole reason underlying group identification. As a vast literature indicates, people are also concerned with acquiring a clear sense of who they are, how they are supposed to behave, and what they should expect from others. Social identification with a group that sends out a clear message regarding these different issues offers an ideal means of reducing subjective uncertainty. In other words, the process of depersonalization with respect to group prototypes that are concentrated, focused, simple, and unambiguous constitutes a most effective way to reduce uncertainty. Such group prototypes are typically associated with groups that are highly distinctive and strongly hierarchical, that is, totalist groups. In support of this conjecture, Hogg summarizes a series of experimental and correlational studies suggesting that an increase in or high level of uncertainty leads people to join groups, and in particular to join groups that are highly entitative, and to identify more strongly with those groups. Such findings, Hogg argues, suggest that the documented success of some totalist groups may well derive from people’s attempts to deal with the increased levels of uncertainty they face in the social environment. Given the impact of a group’s entitativity on perceivers, it comes as no surprise that people may want to keep an eye on the kind of image their own group sends out to others. This is the basic question that Richard L.Moreland and Jamie G.McMinn address in their chapter. The authors start by noting that the current meaning of the word entitativity refers mainly to the perception of a group as being real or not, despite the fact that D.T.Campbell coined the word to refer to the actual rather than perceived reality of groups. Given the contemporary emphasis on perception, the authors propose the concept of “social integration” to refer to the objective reality of a group. Building upon the fact that the levels of entitativity and social integration of groups may sometimes correspond but also diverge, Moreland and McMinn take us on a fascinating analysis of the ways by which people may manage how real their groups is seen to be. A number of different situations may lead people to present their group as more entitative and real than it is. Presumably, declining levels of integration, such as recurrent conflicts within a group, be it a large social entity such as a company or a political party or a smaller group such as a couple, may lead some of its members to exaggerate the strength of the group. As the chapters by Castano and Hogg would suggest, an image of a healthy group may also help to gain new members. Moreland and McMinn review a series of cases in which people may instead want to make the group seem less real that it actually is. Indeed, bot dyadic associations, such as extramarital affairs, office romances, or homosexual relationships, and groups involving larger numbers of people, such as subversive groups, provide evidence aplenty that people are not always keen to let others notice the existence of links among them. Interestingly, although efforts may be made on the part of group members to manage the impression others may have regarding their group, Moreland and McMinn also consider the possibility that group members are sometimes less concerned with the perception of the groupness than with the actual level of social integration of their group. Ironically, a prime strategy here may well be to rely on entitativity. The fact that human beings are social creatures by nature, the authors note, should indeed make it more easy to persuade observers about the presence than the absence of a group. This final chapter confirms, if this was still needed, that entitativity is a most powerful determinant of people’s reactions to social targets. FINAL THOUGHTS Over the last decade, scholars of intergroup relations have made impressive progress in their understanding of the antecedents and consequences of perceived group variability. More recently, this interest generalized to the study of perceived group entitativity and essentialism. Empirical and theoretical advances on the latter two concepts have significantly improved our understanding of important social issues such as the
16
THE PSYCHOLOGY OF GROUP PERCEPTION
development and maintenance of prejudice and discrimination toward groups or the nature of social behavior directed toward (or emerging from) ingroup members and newcomers. As research developed at an ever increasing rate on these notions, it became clear that efforts toward conceptual integration should be made. As we already noted, current evidence suggests that perceived variability, entitativity, and essentialism tend to overlap with each other, both at the conceptual and empirical levels, yet they are also distinct constructs in many ways. Importantly but problematically, scholars in the field have both defined and operationalized these constructs in a variety of different ways. Inevitably, this divergence has meant that research conducted by one laboratory on the relations among these three constructs might be less than useful to other laboratories, where different definitions and operationalizations were used. This distressing state of affairs meant that laboratories were speaking past each other rather than to each other, with the consequence that little conceptual progress was being made through abundant data were accumulating. It was this sad state of affairs that prompted us to edit this volume, with the goal of providing a forum where the diverse points of view could speak to each other. It seemed to us that only through such an exchange could we arrive at some common definitions and perspectives, and that these would lead hopefully to some substantial scientific progress on these concepts. In addition to addressing these theoretical and methodological concerns, we also wanted to document the impressive amount of work that has already been conducted on the role of perceived variability, entitativity, and essentialism in social perception and intergroup relations. We chose to organize these research contributions by focusing on the extent to which the social perceiver is seen to be relatively passive or active in the social construction of group perception. At the one end, perceptions of group variability and entitativity are seen to derive from structural conditions that affect perception and to which the perceiver simply responds. At the other end, perceived variability and entitativity are affected by motivated and even strategic processes that the perceiver actively engages in. Certainly these chapters could have been organized in other ways, along other dimensions. One advantage of the present structure, however, is that it makes clear that the perception of group variability, entitativity, and essentialism can have both negative and positive social consequences. For instance, enhanced perceived entitativity and essentialism facilitate people’s social identification to, or socialization within, groups. However, the search for social identification and integration may occasionally lead people to join totalist groups, or it can also more generally contribute to the development not only of ingroup favoritism but also of outgroup derogation. In this sense, we need to distinguish between the processes involved in perceiving variability, entitativity, and essentialism and their consequences. The processes per se really do not have evaluative implications. That is, perceptions of group variability, entitativity, and essentialism in and of themselves, and the factors that influence these, are neither desirable nor undesirable. But it is the content and interpretation that are put upon these perceptions by the social world that have evaluative implications. Seeing a group as less variable and more entitative is a good thing if one is a member of the group and if one is able to take collective action in pursuit of desirable ingroup goals. On the other hand, outgroups that are less variable, more entitative, and more essentialized are more likely to be seen as effective in acting on their goals and thus, to the extent that there exists conflicts between groups, more likely to be targets of discrimination and intergroup hostility. The various contributions to this book represent a fascinating theoretical and empirical journey into the word of intergroup relations. Our hope is that this volume will stimulate further theoretical and empirical work within social psychology on the fundamentals of group perception. The issues that are raised in this volume extend beyond social psychology, however, and therefore we hope that neighboring disciplines may also find the issues addressed to be provocative. Indeed, one of the obvious lessons of the advances made in the study of group perception is its potential relevance both in terms of concepts and methods for such fields
PERCEIVED VARIABILITY, ENTITATIVITY, AND ESSENTIALISM
17
as political science, sociology, and anthropology. Finally, we also believe that the chapters assembled here will prove useful for social practitioners and decision makers who are generally concerned about social beliefs and interactions. The goal of improving intergroup relations is one that requires efforts from both academics and community-based organizations. We ultimately hope that the present contribution will foster exchanges between these social groups as well.
Section 1 Constructs and Definitions
2 Dynamic Entitativity Perceiving Groups as Actors MARILYNN B.BREWER Ohio State University YING-YI HONG University of Illinois QIONG LI University of Maryland
When D.T.Campbell (1958) introduced the term entitativity to the social science literature, he was interested in the ontological status of social groups as units of perception. The criteria he suggested— similarity, proximity, common fate, (correlation), prägnanz (closure)—were derived from principles of Gestalt psychology and can be understood in two different ways. On the one hand, these criteria can be viewed as properties (actual or perceived) of groups themselves—properties that underlie or legitimate the perception of an aggregate as an entitative social unit. In this spirit, much of the initial empirical research on group entitativity has been devoted to identifying (e.g., Lickel et al., 2000; Harasty, 1996) or manipulating (e.g., Dasgupta, Banaji, & Abelson, 1999; L.Gaertner & Schopler, 1998; S.L.Gaertner et al., 1989; McGarty, Haslam, Hutchinson, & Grace, 1995; Welbourne, 1999) properties of social groups that lead to higher or lower ratings of those groups as “real” social entities. An alternative perspective on Campbell’s criteria for entitativity is to think of them as theories that individual perceivers hold about specific social groups and that endow the groups with social meaning and predictive value (Murphy & Medin, 1985; McGarty et al., 1995). From this perspective, entitativity can be defined as a theory of common origin underlying expected similarities of attitude or behavior on the part of group members. In other words, groups are seen as “real” groups because some aspects of member behavior are believed to arise from some common source. Origin theories provide explanations for group behavior and at the same time afford the boundaries and coherence that make the group a social unit. In this view, entitative social groups are always characterized by homogeneity, similarity, and consistency in some respect, but the question remains, with respect to what? As explanations for behavior, the roots of group entitativity can be further divided into “causes” (distal or proximal) and “reasons” (Malle, 1999). Table 2.1 depicts our taxonomy of origin theories as defined by the nature of the common source underlying group behavior, the type of explanation it represents, and the
20
THE PSYCHOLOGY OF GROUP PERCEPTION
dimension of similarity that would characterize membership in that group. Common attributes are innate, internal dispositions (e.g., genetic characteristics, personality traits, psychiatric disorders) that are shared by all group members and are believed to give rise to similarity and consistency among group members in specific trait-related behaviors or behavioral orientations. When entitativity is based on an implicit theory of common attributes in this sense, intragroup homogeneity is believed to arise from internal, stable causes that are common to all group members. Common history, unlike person attributes, refers to causes of behavior that are external to the individual— past events or experiences that have been shared by all group members, including, for example, a common ancestry, cultural socialization, or life event. Depending on the nature of the shared experience, it may be expected to give rise to similarities in attitudes, values, or physical appearance. In any case, the common history provides a distal (situational) cause for current similarities in (internal) dispositions and associated behaviors.1 Whereas common history refers to past experiences, common fate refers to current or future outcomes that befall all members of the social group by virtue of their group membership. Common fate gives rise to status (or role) similarity among group members that is believed to shape or constrain their behavior. Sociopolitical categories associated with specific social policies (e.g., persons over 65 years old, apartheid categories, labor unions) would be characterized as sharing a common fate, as would groups defined by proximity, such as residents of geographic regions subject to particular meteorological or environmental TABLE 2.1. A Typology of Theories of Group Entitativity Common Origin
Type of Explanation
Domain of Similarity
Common attribute(s) Common history Common fate Common purpose
Person cause Causal history Situational cause Reason
Traits and trait-related behavior Values, attitudes, appearance Status, role Motives, intentions
conditions. Entitativity based on common fate provides an immediate situational cause for commonalities among group members and similarity of behavior. Finally, groups can derive their meaning from common purpose, shared intentions and goals that give rise to collective action. The common purpose provides a reason for group members’ behaviors, which are perceived to be driven by similarity in motives even in the face of diversity of traits, roles, and actions. Common purpose groups are most likely to be characterized by organization and structure (Hamilton, Sherman, & Lickel, 1998), and this is the type of entitativity associated with perceptions of groups as active collective agents with intentions, motives, and goal-directed behavior (Abelson et al., 1998). Implicit theories of common purpose provide a reason (rather than causal explanation) for similarities or consistency in group members’ actions. The different sources of group entitativity depicted in Table 2.1 are by no means mutually exclusive. A specific social group or category can have more than one basis of common origin, and we might expect the different sources to combine additively or multiplicatively to determine the degree of perceived entitativity of the group as a whole. Ethnic groups, for example, are defined primarily in terms of common history/ ancestry, but within many social-political environments they are also characterized by common fate, the perception of innate, common attributes, and the presence of common goals and collective action. Origin theories might also be implicitly correlated. Abelson et al. (1998), for instance, suggest that common attributes may give rise to inferences of common intentions and motives. Nonetheless, although the
DYNAMIC ENTITATIVITY
21
different origins of entitativity may combine and interact, our focus in this chapter will be on the differences among the various types as distinct theories of group entitivity. ESSENCE VS AGENCY: TWO VERSIONS OF ENTITATIVITY The four bases of common origin listed in Table 2.1 have been arrayed along a dimension of “essence” versus “agency” as theories about groups. The distinction between essence and agency is related to the distinction between “entity theory” and “incremental theory” as implicit theories about individual abilities and personal character (Chiu, Hong, & Dweck, 1997; Dweck, Chiu, & Hong, 1995). Entity theorists believe that abilities and personality reflect innate characteristics that are fixed and immutable, whereas incremental theorists believe that the underlying bases of abilities and behavior are malleable and change and develop over time and circumstance. At the group level, essentialist models assume that entitative groups are characterized by a shared “essence” that is fixed, inherent, and immutable (Haslam, Rothschild, & Ernst, 2000; Yzerbyt, Rocher, & Schadron, 1997). The extreme form of essentialism is represented by common attribute categories that are perceived as “natural kinds” (Rothbart & Taylor, 1992). At the other end of the spectrum are theories about entitative groups that emphasize their status as actor-agents. From this perspective, aggregates “become” groups (rather than “are” groups) by virtue of recognizing shared goals and purposes and organizing or banding together for collective action to achieve those goals. Because motives, organization, and action are temporal and malleable (rather than innate and fixed), agency theories of group entitativity correspond to incremental implicit theories at the individual level. Table 2.2 contrasts essence and agency as theories of group entitativity. Drawing on parallels with entityincremental theories of social cognition, essence and agency as theories of group entitativity are expected to set up different processing goals and modes of analysis. Specifically, within the framework of essence theory, perceivers would be motivated to diagnose and evaluate fixed characteristics as the basis for making judgments and decisions about the group and predicting member behavior. In contrast, within the framework of agency theory, perceivers would be motivated to understand the temporal dynamics of group behavior and to predict behavior based on perceived goals in relation to the current situation. With such distinct processing orientations, essence vs. agency theory might orient perceivers to analyze groups differently. Table 2.2 contrasts the analytical processes of essence vs. agency theory on four dimensions. First, in terms of unit of analysis, an essence theory would orient perceivers to focus on personality traits and trait-related units. That is, within this theory framework, groups are perceived and analyzed in terms of the innate, fixed personality traits of their members (e.g., “Blacks are athletic”). Essentialistic theories underlie the formation and use of social stereotypes (Yzerbyt, Rocher, & Schadron, 1997; Yzerbyt, Rogier, & Fiske, 1998). In contrast, an agency theory would orient perceivers to focus on the psychological processes (e.g., goals and values) of group members when they are acting on behalf of the group, which are sensitive to situational changes and account for dynamic group processes. TABLE 2.2. Contrasts between Essence Theory versus Agency Theory Essence Theory Unit of Analysis Personality traits; stereotypes Intragroup similarity Ingroup consistency; look for similarities among group members and consistency of behavior
Agency Theory Psychological processes (e.g., goals, values) Ingroup heterogeneity: look for variations and heterogeneity among group members
22
THE PSYCHOLOGY OF GROUP PERCEPTION
Essence Theory Intergroup similarity Look for clear intergroup distinctions, definition of the boundary of the target group Temporal dimension Look for temporal consistency: e.g., focus on a group’s historical background or traditions
Agency Theory Look for relationships with outgroups; ways to locate the target group in a larger social structure Look for dynamic change: e.g., focus on a group’s future and development
In terms of intragroup similarity, essence theory would have perceivers look for similarities and consistency in behavior among group members as diagnostic of their common underlying attributes. In contrast, agency theory would have perceivers look for variations in behavior and variability among group members as diagnostic of underlying goals and responsiveness to changing situations and functional needs. Coordination, rather than similarity, of behavior is the hallmark of group agency. In terms of intergroup similarity, an essence theory would have perceivers accentuate the differences among groups so as to establish clear group boundaries. With an agency theory, perceivers would be more likely seek to understand the relations among groups and to locate specific groups in a larger social structure rather than as distinct individual units. Finally, in terms of the temporal dimension, essence theory would have perceivers conduct a search into the group’s history so as to locate its common, shared ancestry, history, or traditions. In contrast, agency theory would lead to a search into the future of the group, what it will become (i.e., the development of the group). It should be made clear that we think of essence and agency as two different bases for judgments of group entitativity (even though any one individual may use both theories). This is in contrast to Rothbart and Park (this volume), who posit a single continuum of entitativity (group boundedness) in which common goals are seen as intermediate between common attributes and “essence” in terms of degree of entitivity. In our model, essence and agency reflect two different meanings of the entitativity concept, one based on perceived common attributes, the other based on perceived common goals and intentions. A particular group can be perceived as high or low in entitativity on either basis. Attribute similarity alone does not correspond to high entitativity, unless the attributes have been essentialized. Similarity of goals and purposes alone does not correspond to high entitativity, unless the group is perceived as acting in concert on their common goals. These distinctions between essentialist and agency notions of group entitativity help explain why entitativity judgments are associated only imperfectly with any specific group properties such as similarity or longevity (Lickel et al., 2000). Which group characteristics will be diagnostic will depend on what the underlying basis for judged entitativity is presumed to be. The postulated differences in processing orientations associated with different meanings of group entitativity underscore the role of implicit theories in determining what information about groups is attended to and how group membership is used to explain individual group members’ behaviors and to predict future behavior. The distinction between essentialist and agency theories of group entitativity crosscuts the typological distinctions identified by Lickel et al. (2000) among groups of different size and function (intimacy groups, task groups, social categories, and loose associations). Groups of any type may be viewed in terms of static common properties and similarities or in terms of their dynamic goals and purposes. The judged entitativity of any particular group will depend on which lens is being used. Members of a family, for instance, may be seen as sharing essential features by virtue of heredity and common experience, but if they rarely interact as a group or coordinate their activities, that family will not be seen as a unit in the actor-agency sense. Large
DYNAMIC ENTITATIVITY
23
social categories are usually defined in terms of similarity of characteristics, but nation-states as political units are social actors in a very real sense. In terms of consequences of perceived entitativity, it is our contention that static, essentialist theories of groupness have implications primarily for the perception and evaluation of group members as individuals and of the group prototype (Brewer & Harasty, 1996). These are the processes characteristically studied in social cognition research on stereotypes and stereotyping (e.g., Macrae, Stangor, & Hewstone, 1996). For those who hold essentialist theories of group entitativity, knowledge of group membership provides significant information about the character of individual group members and justifies treating individuals in accord with group stereotypes. For those who hold dynamic agency theories of group entitativity, on the other hand, the nature of groups and group boundaries should be viewed as malleable and changeable over time. A group is a “group” only so long as the collective intent and purpose is salient and shared, and groups can dissolve and re-form as purposes and interests change. Attributions are made to the group as a collective actor, but group membership is not necessarily a meaningful cue to individual characteristics outside of the group context. In this framework, attitudes about the group as a whole may or may not generalize to feelings and evaluations of individual group members. Further, the goals and intents attributed to the group may or may not be linked to essential traits or characteristics of its members. It is possible to attribute aggressive intent to a group without believing that aggressiveness is an essential trait shared by individual group members. Actor-agency theories of groups have more implications for emotional and behavioral responses to the group as a whole (M.G.Alexander, Brewer, & Herrmann, 1999; Brewer & Alexander, 2002). Attributing motives, intent, and purposive behavior to a group as a unit leads to responding to the group as if it were a single individual, but on a larger scale. Given the larger-than-life property of a group acting in concert, it is not surprising that attributions and emotional reactions are magnified in response to perceived entitativity in this sense (Abelson et al., 1998; Dasgupta, Banaji, & Abelson, 1999, Thakkar, 2000). Further, the consequences of perceiving a group as a dynamic entity will vary, depending on whether the group is an ingroup or an outgroup. Agency perception of one’s own ingroup enhances both security and efficacy and hence should be associated with high ingroup identification and positivity (see Castano, this volume). The consequences of outgroup agency, on the other hand, depend on the nature of the relationship between ingroup-outgroup goals, and their relative status and power (Alexander et al., 1999). A trusted outgroup would be more valuable as an ally if it is also perceived as an entitative actor-agent. However, a distrusted or disliked outgroup becomes more threatening and dangerous to the extent that it is perceived as a dynamic entity or collective actor (Dasgupta et al., 1999). Static traits or essences that are attributed to the outgroup may contribute to the initial evaluation of the group as a potential friend or enemy, but it is the dynamic attributes of the group-as-actor that determine whether the outgroup can benefit or harm the ingroup. Thus, the study of the relationship between group cognition and intergroup relations first requires an understanding of whether an essentialist or agency theory of group existence has been activated in any given case. Essentialist notions of group entitativity are covered extensively in other chapters in this volume. The remainder of this chapter will be devoted to further discussion of defining entitativity in terms of agency theories of groups. PERCEIVING GROUPS AS ACTORS: WHEN AND WHY Activating dynamic agency theories of group entitativity is a joint function of permanent and situational properties of the group and of chronic and temporary beliefs and motives of the individual perceiver.
24
THE PSYCHOLOGY OF GROUP PERCEPTION
Certainly many of the properties identified by D.T.Campbell (1958) as cues that perceivers might use in assessing whether an aggregate of individuals constitutes a meaningful group could be considered indicative of common purpose and intent. In particular, proximity and covariation of movement in time and space are dynamic properties that elicit perceptions of the group as a coordinated, acting unit (Dasgupta et al., 1999). However, outside of the behavior of sports teams and angry mobs, coordination is not always so visibly evident as a property of groups. Groups can share purposes and organize action even at a distance, with or without formal communication networks. Thus, the dynamic properties of shared intention and coordination may be psychological states and relationships among group members rather than formal organizational structures. As such, the degree of dynamic entitativity perceived may be as much a product of the beliefs and motives of the perceiver as of the reality of the group itself (Brewer & Harasty, 1996). In the political domain, the portrayal of outgroup nations or political parties as monolithic actors with malevolent intent and organized purpose is often used strategically to mobilize ingroup cohesion, loyalty, and collective action. There is experimental evidence that this strategy is effective in that manipulation of the perceived entitativity of an outgroup is sufficient to enhance perceptions of the outgroup’s power and resources and to increase perceived ingroup entitativity and the level of ingroup identification (Castano, Sacchi, & Gries, in press). The use of this strategy in intergroup contexts may have self-fulfilling effects. If outgroup entitativity increases perceived cohesion and identification with the ingroup, it creates the conditions under which members of the ingroup are most likely to behave as a collective actor with a common purpose. Apart from strategic motives, implicit theories about the world (Dweck, Chiu, & Hong, 1995; Chiu, Dweck, Tong, & Fu, 1997) may increase the likelihood that social groups are perceived as actor-agents with intents and purposes. Specifically, to the extent that individuals perceive the social world and its institutions as malleable, they are likely to focus on how social groups change and develop and so imbue social groups with the properties of active agents. In contrast, to the extent that perceivers hold the view of a fixed social world, they may focus on identifying the static attributes that define the groups (such as the common attributes, common history, and common appearance among members) rather than the dynamic aspects (such as common purpose). This tendency to endorse the essentialist theory of group entitativity should be especially intense if the fixed worldview is coupled with a belief in low efficacy of social groups in causing important outcomes (i.e., low group efficacy belief). In this case, the fixed world may be seen as consisting mostly of social groups that are formed according to traditions or common practices, but seldom function as effective actor-agents. The role of the individual in this system might be to navigate a way through these fixed structures and get the best out of them. Previous research has also shown evidence of cross-cultural differences in the endorsement of the fixed vs. malleable theories of the world. Specifically, Chiu and Hong in several studies have consistently found that the Chinese samples hold a more fixed (vs. malleable) belief about the world than did the American samples (Chiu & Hong, 1999; see review in Levy, Plaks, Hong, Chiu, & Dweck, 2001). These crosscultural differences in implicit theories of the world have been shown to link to American people’s tendency to shape social structures to accommodate individuals’ needs and Chinese people’s tendency to accommodate to the social structure (Su, Chiu, Hong, Leung, Peng, & Morris, 1999). In perceiving social groups, American participants might also endorse the actor-agent theory of group entitativity. In contrast, Chinese participants might endorse the essentialist theory of group entitativity, especially those who also think that the social groups have low efficacy in causing important outcomes (i.e., low group efficacy belief). These ideas were tested in a study that we will discuss next.
DYNAMIC ENTITATIVITY
25
DYNAMIC ENTITATIVITY ACROSS CULTURES: A PRELIMINARY INVESTIGATION As an initial exploration of possible cultural differences in the perception of groups as entities—particularly as actor-agents—we undertook a survey study among college students in Hong Kong and in the United States. During the same time period at the beginning of the fall 2000 academic term, 101 students at the Hong Kong University of Science and Technology and 246 students at the Ohio State University completed questionnaires assessing their perceptions of specific social groups to which they belonged.2 The target groups included in the survey questionnaires were selected to represent a range of group size and type. Specifically, the groups chosen for investigation were (1) immediate family (intimacy group), (2) a social club or other small group affiliation (intimacy or task group), (3) university (task group or social category), and (4) nation (social category or loose association). Individual respondents completed the full questionnaire for two of these groups (either family and university or small-group and nation), with order counterbalanced. The questionnaires were written in Chinese for the Hong Kong sample and in English for the U.S. sample. Entitativity Measures Sets of items in the questionnaire were written to assess three different entitativity-related aspects of group perceptions. One set of items assessed dynamic properties; a second set assessed relatively static properties and essentialist beliefs; and the third measure was a judgment of the general “groupness” of the group as a whole. Static Characteristics and Essence. The perception that members of a group have “common attributes” is one basis for group entitativity (see Table 2.1). However, high attribute similarity does not necessarily imply high perceived entitativity in and of itself, but only when these attributes are seen as fixed, innate characteristics that reflect the “core essence” of the group. In order to tap this essentialist belief directly, two items in the questionnaire assessed the extent to which the group was perceived as having such fixed characteristics and a core essence. These two items constituted our measure of perceived essence ( =.68). Dynamic (Agency) Characteristics. Each group was also rated by selected dynamic properties, including common fate (future), shared purpose, group cooperation, and effectiveness as a group. These four items were intended to represent the bases of group entitativity associated with an actor-agent theory of entitativity. Ratings of each group on these five items were averaged together to form our measure of perception of the group-as-actor ( =.82). Entitativity Rating. Finally, two items were included in the questionnaire to create a general measure of perceived entitativity of the group. These items were (1) the extent to which the group should be “thought of as a whole” (rather than a collection of individual members), and (2) the extent to which the group is a coherent unit ( =.82). All ratings were made on an 11-point scale, ranging from 0 to 10, with 5 as the midpoint. Implicit Theories of the Social World and Group Efficacy Beliefs In addition to the measures of target groups, the questionnaire included a measure of the implicit theories about the social world developed by Dweck et al. (1995). A revised version of the measure consisted of four items that depict a fixed view about the world (e.g., “Though we can change some social phenomena, it is unlikely that we can alter the core characteristics of our social world.” “Our social world has its basic or ingrained characteristics, and you really can’t do much to change them”), and four items that depict a
26
THE PSYCHOLOGY OF GROUP PERCEPTION
malleable view about the world (e.g., “We can change some social phenomena, and even the core characteristics of our social world.” “Just as societal trends can change, so can the fundamental nature of our social world”). Participants were asked to rate the extent of their agreement with the statements from 1, not at all, to 6, very much. The internal reliability of the scale derived from the eight items (with malleability items reverse scored) was =.83. The efficacy scale (developed by Hong and Wong, 2000) consisted of 12 items asking the respondents’ agreement with statements such as “The will of individuals is the most powerful force in society,” “Social groups and organizations influence what happens in an individual’s life,” “What happens in an individual’s life is of his or her own making,” and “Social groups and organizations take control of the situations around them and exercise free will.” Six of the items referred to individual will and influence, and six of the items referred to group will and influence, creating individual efficacy and group efficacy subscales, respectively. These measures were intended to tap chronic individual beliefs that were hypothesized to be related to perception of groups as static or dynamic entitities. Cultural Differences in Perceiving Groups as Entities: Some Findings Our first analyses of the survey data took a look at mean differences on the three entitativity-related measures as a function of respondent culture and target group, and aggregated across all four target groups within each culture. The relevant means are reported in Table 2.3. As is evident from inspection of the table. Hong Kong respondents consistently rated the target groups lower in overall entitativity than did U.S. respondents. Across all target groups, the main effect of culture on entitativity ratings was statistically significant. However, given issues of comparability due to translation and possible differences in the use of response scales, interpreting mean differences across cultures in an absolute sense is tenuous at best.3 Of more interest are differences between cultures in the relative ratings on the different measures, as well as differences within cultures in the perceived entitativity of the various target groups. In both cultures, large social categories (university and nation) were rated as lower in entitativity than the family group. However, Hong Kong respondents made a sharp distinction between the family as a rela TABLE 2.3. Mean Entitativity Ratings by Culture and Target Group Hong Kong Family Small Group University Nation Aggregated United States Family Small Group University Nation Aggregated
Entitivity
Dynamic
Essence
6.93 6.24 6.06 6.04 6.32
6.00 6.08 5.65 5.78 5.88
6.94 5.95 6.24 6.31 6.36
7.69 7.70 6.68 6.67 7.18
7.36 7.41 6.72 6.20 6.92
7.53 6.90 6.63 6.78 6.96
DYNAMIC ENTITATIVITY
27
tively high entitative group and all other types of groups, including small social clubs. U.S. respondents, on the other hand, rated the two types of small, interactive groups essentially the same, and relatively high on entitativity and dynamic properties especially. Of more interest to our purposes than mean ratings of entitativity were the relationships among the entitativity measures within and between cultures. More specifically, we were interested in the extent to which perceptions of a group in terms of essence or actor would predict the overall entitativity rating of that group. For each target group within each culture, multiple regression was used to assess the relative weight of dynamic (actor) versus static (essence) group TABLE 2.4. Multiple Regressions: Predictors of Entitativity Ratings Hong Kong Family Small Group University Nation Aggregated United States Family Small Group University Nation Aggregated Note: *=p<.05 **=p<.01
Actor ( )
Essence ( )
R2
.58* .37** .56** .47** .49**
.27* .42** .23* .35** .33**
.56** .41** .42** .54** .48**
.51** .32** .48** .33** .45**
.34** .27** .21* .44** .32**
.62** .25** .39** .41** .46**
ratings as predictors of the group rating on the general entitativity scale. In addition, an overall regression analysis was conducted on the data for each culture aggregated across all four target groups. Table 2.4 reports the results of these regression analyses. Although there were some differences from group to group, overall the data on prediction of entitativity looked remarkably similar between the two culture samples. In both cultures, essence ratings and actor ratings made significant independent contributions to predicting group entitativity.4 Further, in most cases, the dynamic actor model carried somewhat more weight than did the essence model in predicting overall entitativity, and this was true across both cultures. Although the cultures did differ as predicted in their relative beliefs in a fixed versus malleable social world (the mean score of Hong Kong participants on the implicit theory of the world measure, M=4.08, was significantly higher than that of the U.S. participants, M=3.64, indicating that the Hong Kong participants endorsed a fixed theory of the world to a greater extent than did the U.S. participants), this difference in worldview did not produce any differences in the relative weight of essence or actor-agency contributions to entitativity scores either within or between cultures. The Role of Group Efficacy Beliefs The significant contribution of both essence and actor ratings could mean either that both static and dynamic theories of entitativity are used to make judgments of the overall coherence of the group by all
28
THE PSYCHOLOGY OF GROUP PERCEPTION
individuals, or that different individuals rely on one theory or the other. To assess the latter possibility, we looked at the role of individual differences in efficacy beliefs within cultures. Overall, the mean of Ohio students on both individual efficacy (M=6.89) and group efficacy (M=6.69) was higher than that of Hong Kong students (M =5.87 and 5.81, respectively). Nonetheless, there was equivalent variability within both cultures on beliefs in group efficacy relative to individual efficacy. To control for overall mean differences, the difference between individual efficacy and group efficacy scores was computed for each respondent and used as an index of relative group efficacy belief within both cultures. A median split (within culture) on this index was used to identify those relatively high in group efficacy belief and those relatively low in perceived group efficacy. This subdivision was used in further analyses to explore the moderating role of implicit beliefs on the determinants of group entitativity ratings. Multiple regression analyses of entitativity scores were recomputed separately for those high and low in group efficacy within each culture, using the aggregated data. The results of these analyses are reported in Table 2.5. For the American sample, the resulting beta-weights were virtually identical for both subgroups. Dynamic actor ratings predicted group entitativity ratings in either TABLE 2.5. Predictors of Group Entitativity: Individual Differences in Hong Kong and in the United States Hong Kong Low group efficacy High group efficacy United States Low group efficacy High group efficacy Note: *=p<.05 **=p<.01
Actor ( )
Essence ( )
.50** .52**
.44** .18 (ns)
.41** .48**
.35** .30**
case, with essence ratings also making a significant contribution. With high efficacy beliefs overall in the United States, it apparently made little difference whether individual efficacy was seen as stronger than group efficacy or vice versa. For the Hong Kong sample, however, individual differences in beliefs about locus of efficacy did moderate the relative importance of dynamic and static group properties in determining entitativity judgments. Respondents who had relatively low beliefs in group efficacy showed a pattern similar to that in the United States, with both models of group entitativity being significant predictors. However, those respondents in Hong Kong who were relatively high in belief in group efficacy beliefs consistently based their entitativity judgments primarily on their perceptions of dynamic properties of the group with no significant contribution of essence ratings, suggesting a group-as-actor view of group entitativity. Thus we find some evidence that in a culture where fixed worldview dominates, low efficacy beliefs are associated with more reliance on essence theory of entitativity than are high efficacy beliefs. However, there was no group for which essence ratings contributed more than dynamic actor ratings to the general entitativity measure. This apparent dominance of the dynamic agency view of group entitativity across both of our culture groups may be due in part to the fact that respondents in our study were making evaluations of their own ingroups. Research on self-knowledge and social inference (e.g., S.M.Andersen & Ross, 1984) indicates that individuals tend to think about others primarily in terms of behaviors and traits but are more likely to think about themselves in terms of feelings, goals, and intentions. This tendency to focus self-perception on intentions and motives rather than attributes may also be extended to salient ingroups (as extensions of the
DYNAMIC ENTITATIVITY
29
self). Thus it is possible that cultural differences in underlying theories of entitativity maybe moderated by whether ingroups or outgroups are being evaluated. CONCLUSION The initial results of our comparison between Hong Kong and the United States provide at least preliminary support for our contention that essentialism and dynamic agency representations of groups represent distinct views of group entitativity that contribute independently to perceptions of a group as an entitative unit. For purposes of theory testing, we focused here on individual and cultural differences in perceptions of groups as static essences or dynamic actors. However, the data made it clear that these views of group entitativity are not mutually exclusive and that individual perceivers can assess the meaningfulness of social groups in terms of both theories, either simultaneously or at different times. Thus, both views coexist, and it may be of value in future work to specify more clearly what form of group entitativity is being assessed or activated when individual respondents describe a group as a meaningful unit. NOTES 1. Note that different theories of the meaning of ethnicity or ethnic/racial groups provide a good example of some of the distinctions involved here. For some, ethnicity represents a biological category involving shared genetic heritage that gives rise to common attributes and traits. For others, ethnicity is a cultural heritage, reflecting shared history and common socialization experiences. 2. Since part of our interest in this study was the relationship between social identification and perceived entitativity, all the target groups in this survey were ingroups. 3. Overall, the American sample averaged higher ratings on almost all the response scales in the survey, so any mean differences have to take into account this possible response style artifact. 4. In every regression, the combination of both variables increased the R2 significantly over the regression based on either predictor taken singly. In addition, when we added a multiplicative term representing the interaction of essence and actor ratings, this term did not contribute significantly to an increase in the R2 in any case. The two underlying bases of entitativity were clearly independent and additive in their effects.
3 Perceiving the Groupness of Groups Entitativity, Homogeneity, Essentialism, and Stereotypes DAVID L.HAMILTON University of California, Santa Barbara STEVEN J.SHERMAN Indiana University JULIE S.RODGERS University of California, Santa Barbara
Research on group perception has a long and varied history in social psychology. One of the earliest topics to be studied empirically was the nature of people’s stereotypes about a variety of national and ethnic groups (D.Katz & Braly, 1933). The study of stereotypes has continued to be an active area of inquiry ever since and has produced a massive literature exploring numerous aspects of the belief systems people generate and adhere to in their conceptions of groups (G. Allport, 1954; Brigham, 1971; Fiske, 1998; Hamilton & Sherman, 1994). In this history we have learned a great deal about the cognitive and motivational underpinnings of stereotype formation, use, and preservation (cf. Macrae, Stangor, & Hewstone, 1996). Certainly the study of the categorization process, as a fundamental cognitive mechanism, has been central in the literature ever since Allport’s (1954) classic text highlighted its role in “the normalcy of prejudgment.” Of course, the burgeoning research literature stimulated by social identity theory (Tajfel & Turner, 1986) and self-categorization theory (Turner, Hogg, Oakes, Reicher, & Wetherell, 1987) has placed the categorization process on center stage in analyses of group stereotypes and has highlighted the motivational benefits that can be gained by intergroup differentiation. Recently the study of group perception has turned to focus on some more basic questions that precede, and perhaps accompany, the formation of group stereotypes. Specifically, researchers have focused their attention on questions concerning when and how people come to perceive groups as groups, and this line of inquiry has expanded into a variety of offshoots that examine different, though related, issues in group perception. In this chapter we discuss three of those lines of work—research on perceptions of group entitativity, group homogeneity, and group essentialism. Whereas several other chapters in this volume focus on homogeneity and essentialism, our discussion centers on our own and others’ recent research on perceptions of group entitativity. We review our findings, and we consider the ways that we and other researchers have used the concept of entitativity, both conceptually and empirically. We are particularly
PERCEIVING THE GROUPNESS OF GROUPS
31
concerned with the interrelations among the three concepts, as well as the implications of these concepts for stereotypes and stereotyping. THE FOUNDATIONS OF GROUP PERCEPTION An interest in the fundamental building blocks of group perception is not new. For the last two decades researchers have intensively researched the nature and role of social categorization in intergroup perception, and the importance of categorization and its consequences is now well established. Once people have been categorized into a group, however, the perceiver may (or may not) endow that category with other features and properties associated with “groupness,” and may do so to varying degrees. Research during the last few years has shed new light on what it means to perceive an aggregate of individuals as a group, the qualities that contribute to that perception, the different types of groups that perceivers routinely distinguish in everyday perception, and the consequences that follow from endowing groups with these properties. Extending the bases of group perception beyond categorization, three concepts have been discussed as playing important roles in group perceptions, and each of them has implications for how we think about stereotypes. In this section we briefly summarize each of these concepts and their implications for stereotyping. Homogeneity A commonly observed finding in the literature on intergroup perception is that perceivers judge members of an outgroup to be more homogeneous, to have less variability, than members of their own ingroup. For many years this finding was regarded as a truism, and various theoretical accounts—emphasizing both cognitive and motivational processes—were proposed and debated as explanations for this presumably basic and robust aspect of group perception (Devos, Comby, & Dechamps, 1996; Doosje, Spears, Ellemers, & Koomen, 1999; Linville & Fischer, 1993; Linville, Fischer, & Salovey, 1989; Messick & Mackie, 1989). One reason for the great interest in perceptions of outgroup homogeneity is its obvious relevance to stereotyping. One of the hallmarks of stereotyping is overgeneralization (G.Allport, 1954), in which stereotypic attributes are routinely applied to a wide range of people (all members of a category), making them all functionally equivalent to each other in the perceiver’s mind’s eye. Consequently, conditions that foster perceptions of outgroup homogeneity— “They’re all alike”—would seem to make it easier to form and apply stereotypic concepts regarding the group. Presumably, then, there should be an inverse relation between perceptions of group variability and stereotyping of that group (Park & Hastie, 1987; Ryan, Judd, & Park, 1996). Thus, based on perceptions of outgroup homogeneity, there should be greater stereotyping of outgroups than of one’s ingroup. However, things are never quite as simple and direct as they seem. After several years in which the outgroup homogeneity effect had the status of a truism, researchers began to generate instances of the opposite outcome, that is, demonstrations of an ingroup homogeneity effect—greater perceived homogeneity in the ingroup than in the outgroup (Ryan & Bogart, 1997; Simon, 1992a; Simon & Brown, 1987). Theoretically, this finding raised a number of conceptual challenges. Under what conditions are group members likely to perceive themselves as being highly homogeneous, even more so than some alternative group? Why? And when this happens, does this imply that group members apply stereotypic conceptions to their own group? What we are considering here, of course, is exactly what Turner et al. (1987) have proposed as self-stereotyping, that is, perceiving oneself as interchangeable with other ingroup
32
THE PSYCHOLOGY OF GROUP PERCEPTION
members and therefore adopting, in one’s self-conception, those attributes that are stereotypic of the ingroup. As documented in several of the chapters in this volume, the evolution of this research literature has generated a series of conceptual and empirical advances that have shed considerable light on some central questions about intergroup perception. Nevertheless, important questions remain to be answered. For example, we need to know more about exactly when the two opposing effects—ingroup homogeneity and outgroup homogeneity—occur and what factors drive the outcome one way or the other. We also need to know more about the extent to which perceptions of group homogeneity do in fact enhance the tendency to overgeneralize about the group’s members and to endow them with group-defining stereotypic attributes, and whether this happens in comparable degrees for ingroup and outgroup judgments. There have been several approaches to addressing this issue concerning when attributes generalize across various members of a group and when they do not. For example, Rothbart and Lewis (1988) reported that the tendency to generalize a behavior or an attribute from one target member of a group to other members was a function of the prototypicality of the target member. In other words, the more similar the target member is to the average member of the group, the more likely the tendency to generalize the traits of that member to the group as a whole and to other members. At a more general level, Brewer (1988) and Fiske and Neuberg (1990) have discussed the broad question of when social perceivers view group members in terms of their unique individuated characteristics and when they view them in terms of the categorical attributes of their group, generalizing from the qualities of the group as a whole or from the attributes of other members. Essentialism Categorization theorists have recently proposed that people distinguish between two qualitatively different kinds of categories, typically called natural kinds and human artifacts. Natural-kind categories are believed to reflect some underlying essences, often biological in nature, that clearly differentiate one category from another, whereas human-artifact categories represent the products of human ingenuity and do not typically possess some inherent quality or biological basis. Such categories are defined by function rather than underlying essence. Rothbart and Taylor (1992) have made the argument that social perceivers regard certain human groups as natural-kind categories, implicitly assuming that there is some underlying essence that bonds members of that category together, and they proposed that stereotypes of some groups may reflect such essentialistic beliefs. More recently, Yzerbyt, Rocher, and Schadron (1997) have developed this viewpoint further, particularly with regard to its application to understanding stereotypes. The essentialism argument is that perceivers endow certain groups, most typically large social categories, with the properties of natural kinds, despite the fact that by any criteria involving a common essence it is hard to regard them as natural kinds. One of the important consequences of viewing a category as a natural kind is that it permits an extensive set of inferences to be drawn, such that natural-kind categories have greater inductive potential than do human-artifact categories. This enhanced potential for inferences about inherent qualities reflects the oftenheld belief that surface-level cues that identify category membership directly reflect the underlying essences that bind those members together in a meaningful group. When applied to human groups, this inductive potential might allow the overgeneralization that is so commonly associated with stereotyping. A second property of natural-kind categories is unalterability, the belief that members of such categories cannot easily change membership into some alternate group. This property again reflects the inherent assumption that the category is based on an underlying biological essence defining the group, an essence that is not easily changed. Hence, unalterability implies that once a member of the category, always a member of the
PERCEIVING THE GROUPNESS OF GROUPS
33
category. Again, when applied to human groups, this feature facilitates the belief in enduring characteristics, in that membership in the group is unlikely to change. In order to test some of the assumptions of the essentialistic framework, Haslam, Rothschild, and Ernst (2000) had participants rate 20 different social categories on a variety of measures that represented properties attributed, by various authors, to natural-kind categories, including the two cited above. Interestingly, their factor analysis of these properties identified two independent factors representing properties of natural-kind categories, and in fact inductive potential and unalterability appear to capture these different factors. One of Haslam et al.’s factors reflected the beliefs that the category has some “underlying reality,” that the members are highly similar, that such membership excludes membership in certain other categories, and the extent to which membership is informative about category members (inductive potential). The second factor included measures tapping the perceived “naturalness” of the group, the belief that there are certain “necessary characteristics” for membership in a category, the immutability (unalterability) of group membership, the discreteness of membership, and the stability of the group over time. Entitativity Groups are collections of individuals, but not all collections of individuals are groups. For example, 100 persons carrying placards and clubs, steadily moving along a street toward City Hall, are considered a group (protestors, demonstrators, rioters), whereas 100 persons, carrying briefcases, hurriedly moving through Grand Central Station from train to street, would not usually be considered a group (business people individually rushing to their offices). On what bases does the observer of these aggregates decide that one is a group and the other is not? When the “groupness” of a collection of persons is perceived, are all such groups perceived in the same way, or does the perceiver distinguish different types of groups, and if so, on what bases? Moreover, once such a collectivity is perceived to have the quality of being a group, what difference does it make for the perceiver’s subsequent perceptions? And if a group is perceived to be a solid, unified group, one that constitutes the kind of real “entity” about which strong beliefs might easily develop, would that mean that stereotypes are most likely to form about groups that are perceived to be high in entitativity? Conversely, by the same reasoning, are groups about whom we have stereotypes more likely perceived to be high in entitativity? Although these questions are not entirely new (D.T.Campbell, 1958), they have only recently become the focus of research and analysis (Abelson et al., 1998; Brewer & Harasty, 1996; Hamilton & Sherman, 1996; Hamilton, Sherman, & Castelli, 2002; Hamilton, Sherman, & Lickel, 1998; Hamilton, Sherman, & Maddox, 1999; S.J.Sherman, Hamilton, & Lewis, 1999). These questions, and recent research findings relevant to them, are discussed at length in later sections of this chapter. The “Easy” Blending of Concepts These three concepts would appear to be intimately related to one another. For example, as D.T.Campbell (1958) indicated, the similarity (homogeneity) among a groups members can serve as an important stimulus feature the perceiver might use as a cue to the group’s entitativity. The converse relation would also seem to be true: If a collection of people is perceived to be a meaningful, highly entitative group, then it seems plausible that those individuals would be perceived as sharing similarities that might induce perceptions of homogeneity in the group. Moreover, if that group is perceived as a coherent entity with homogeneity among its members, then the assumption that there is some “essence” underlying the group, some latent quality (perhaps even biological in origin) that affords this unity, becomes quite powerful, and hence the group is
34
THE PSYCHOLOGY OF GROUP PERCEPTION
seen as all the more “real.” Once viewed as a natural kind, the group is then seen as stable and its membership unalterable, and its consequent high inductive potential permits the inference of a broad and extensive array of features and attributes, the very stuff of which stereotypes are made. Given this reasoning, it appears that these concepts have provided a considerable advance in our understanding of stereotypes. And perhaps this is true. However, as impressive as the research associated with these concepts has been (as is well evidenced in the chapters of this volume), the easy blending of concepts reflected in the preceding paragraph is, we feel, premature. Hence our message in this chapter is that things are more complex than they may seem, and any sense of complacency and closure is not warranted. More specifically, we argue that these concepts refer to distinct aspects of group perception and therefore any serious “blending” of the concepts is unwarranted. The arguments we develop in this chapter derive primarily from our own research on entitativity, and therefore we develop those points in the context of discussing research on entitativity and its relations with other concepts. STUDYING ENTITATIVITY: ISSUES OF CONCEPTION, MANIPULATION, AND MEASUREMENT D.T.Campbell (1958), who first introduced the term entitativity more than 40 years ago, defined it simply as “the degree of being entitative. The degree of having the nature of an entity.” Although this definition lacks the kind of specification we might desire, researchers since Campbell have rarely offered conceptual clarification, often merely citing his definition as their guideline. Within the last decade several authors have attempted to develop this concept, relating it to other ideas in the literature. For example, Brewer and Harasty (1996) developed a correspondence between the perceived entitativity of a group and the extent to which that group is represented, in the perceiver’s mental structures, by a prototype. In particular, they argued that minority groups are more likely to be represented by prototypes than are majority groups, from which thesis they derived the prediction that minority groups will be seen as possessing more entitativity than will majority groups. A somewhat different argument has been advanced by Yzerbyt, Rocher, & Shadron (1997; see also Yzerbyt, Rogier, & Fiske, 1998), who linked entitativity to the similarity among group members as well as to notions of essentialism underlying stereotypic conceptions of social categories. Hamilton et al. (1998) discussed several bases of perceived entitativity and emphasized the perception of organization and structure in the group as a primary cue to this quality of groupness. Although all of these conceptions can be viewed as extensions and developments of ideas first posed by D.T.Campbell (1958), there is little empirical basis for differentiating among them or for favoring one over another. Given this uncertainty, it perhaps is not surprising that the very concept we study lacks precision. As one step toward such clarification, Hamilton et al. (2002) defined entitativity as “the perception that an aggregate of individuals is bonded together in some way to constitute a group.” Of course, it is the basis for that “bonding together” that needs amplification and specification. For any given group, what is the nature of the “social glue” that holds them together in a group? This is a difficult question, and various authors have emphasized different factors on which “groupness” is based. In our view, the answer to this question will in all likelihood differ for different groups. That is, people may be bonded together by different kinds of “glue” in different kinds of groups. In this chapter we hope to elucidate this question and thereby both expand and constrain what we mean by entitativity.
PERCEIVING THE GROUPNESS OF GROUPS
35
Cues to Entitativity The lack of clarity and consensual meaning of the term entitativity has also presented problems for efforts to study this concept. Perhaps the major contribution of D.T.Campbell’s (1958) conceptual analysis of entitativity was in specifying a number of cues that might be used by perceivers to infer whether a collection of individuals qualifies as a group or is merely an aggregate. Thus, if a perceiver, observing a collection of individuals, notices that they are in close proximity, share common goals or common fate, function in coordinated fashion, appear to have impermeable boundaries, or appear similar to each other in some way, then those attributes will increase the likelihood of perceiving those individuals as comprising a group. Remarkably, until recently there has been very little research exploring the implications of these various cues for perceiving group entitativity. The approach we adopted to studying entitativity was initially derived from Campbell’s intuitive analyses of group perception. Specifically, we sought to determine empirically the extent to which each of several cues—those cited by Campbell, plus several others—were empirically related to perceptions of the “groupness” of a broad array of groups that we encounter in everyday life. Lickel, Hamilton, Wieczorkowska, Lewis, Sherman, and Uhles (2000) had subjects rate a broad array of 40 different groups (e.g., members of a family, students at a university, a company committee, women, a jury, people who live in the same neighborhood, Jews) on a variety of rating scales. These scales included a rating of the extent to which the group “qualifies as a group,” which served as our measure of perceived entitativity. They also rated the 40 groups in terms of the extent to which members interact with each other, the importance of the group to its members, the extent to which they share common goals, the extent to which they experience common outcomes, the similarity among group members, the size of the group, the length of its history (duration), and the ease or difficulty of joining or leaving the group (permeability). We found that the first five of these variables—interaction, importance, common goals, common outcomes, and similarity—were most strongly correlated with our “groupness” ratings, whereas size, duration, and permeability were essentially uncorrelated with entitativity across these 40 groups. One implication of these findings, in light of our previous discussion, is that the degree of similarity (homogeneity) among group members may be related to perceived group entitativity, but it is only one of several such properties that may be diagnostic of groupness, as judged by perceivers. Types of Groups The list of 40 groups rated by Lickel et al.’s (2000) subjects included a diverse array of social groupings, and the findings document that they span a considerable range on an “entitativity continuum” (Hamilton et al., 1998). Beyond this variability in groupness, it may be that there are meaningful distinctions among different types of groups, and these types may also differ in their entitativity values. We therefore conducted some further analyses to explore that possibility. In these analyses, the eight ratings of group properties (excluding the entitativity measure) were treated as a “profile” of each group, and these ratings were then analyzed to identify clusters of groups whose ratings profiles resembled each other (and differed from those of other clusters). This analysis identified several meaningfully distinct “types” of groups that differed in entitativity-related properties. Intimacy groups (family, a group of friends, a street gang, a support group) are small, are highly interactive, are important to their members, are generally impermeable, and have been in existence for a long time. Task groups (a committee, students studying together for an exam, a jury, the cast of a play) are also small and interactive, are not as important to its members, are of much shorter duration, share common goals, and are much easier to join or leave. Social categories (women, blacks, Jews, Americans) are very large groups that are low in interaction, are moderately important to members, are highly
36
THE PSYCHOLOGY OF GROUP PERCEPTION
impermeable, and have long histories. Finally, loose association groups (people living in the same neighborhood, people who like classical music, students at a university) are moderate in size and not very interactive, are of low importance to members, and are highly permeable. Two additional points are noteworthy. First, no single group feature consistently distinguishes one type from another. Rather, each type is defined by a pattern of features or properties, and those patterns differ meaningfully. Second, these four group types also differ significantly in the extent to which they are rated as being entitative. Intimacy groups are highest in perceived entitativity, followed by task groups, then by social categories, and finally by loose associations. These clusters represent four qualitatively different, empirically derived group types that reflect distinctions that perceivers make in sorting the various groups that populate their social world. Essentially the same group types emerged in parallel studies conducted in two countries, using rating scale and sorting techniques, and in analyses based on groups to which the perceiver belonged or groups to which the perceiver did not belong (see Lickel et al., 2000, for details). Moreover, these group types are not simply a product of deliberative rating or sorting tasks, as more recent evidence has shown that this typology is spontaneously used by perceivers in processing and representing information about group members in memory (S.J.Sherman, Castelli, & Hamilton, 2002). In this research, participants were shown a series of pictures of faces, along with a label indicating the group membership of the person portrayed. Six different group labels were used, two for each of the major group types (intimacy groups, task groups, and social categories). In a subsequent recognition task, only the faces were presented, and participants were asked to remember the group membership of the person. Results indicated significantly more within-group-type errors than between-group-type errors. For example, a member of a Presbyterian group (a social category) would more likely be recalled as a member of a French group (another social category) than as a jury member (a task group) or a family member (an intimacy group). These results reveal the spontaneous use of an implicit group typology and its effect on the cognitive organization of information about social groups. The accumulated findings, then, give us confidence that this typology reflects some rather fundamental distinctions in the social perceptions of everyday perceivers. These findings also illustrate, in a potentially useful structural typology, the immense variety of entities to which we routinely apply the word group. The Uneasy Tensions Among Concepts The distinctions among different group types reflected in Lickel et al.’s (2000) findings have important implications for some of the concepts discussed in the first section of this chapter. First, note that social categories—blacks, women, Italians, gays—constitute only one of several types of groups, and note also that those social categories received only intermediate ratings on the “qualifies as a group” measure. There are other types of groups that are perceived as considerably more unified and coherent as groups, yet both lay people and researchers rarely talk about stereotypes of these other groups. Why? One answer to this question is suggested by self-categorization theory. Social categories are often thought of in the context of some contrast group(s)—men/women; black/ white; Jew/Catholic/Protestant. When a group is viewed in the context of a contrast group, it would seemingly both enhance the perception of it as an entitative group and increase the likelihood of stereotype formation and use. Of course, as emphasized by self-categorization theory, the specific contrast group(s) engaged can vary from one context to another and therefore, as Abrams (1999a) has pointed out, the extent to which a social category is perceived to be a meaningful unit can be flexible and context-dependent. By implication, the perceived entitativity of a group might also fluctuate with changes in this comparative context. These effects of contrast groups seem less likely in the case of intimacy (family, friends) and task (juries, labor unions) groups.
PERCEIVING THE GROUPNESS OF GROUPS
37
Second, it is interesting that social categories—about whom much work on perceptions of group homogeneity has been invested—received fairly low ratings on the question of how similar members of the groups are to each other. One might expect that this is the variable on which social categories would be rated highly; after all, we even have colloquial sayings about such similarities (“They all look alike.” “You’ve seen one of them, you’ve seen them all”). These statements imply widespread similarity among members of social categories. However, there is an important distinction here. It is true that, for many social categories, there is some feature on which all members of the category are similar, often the attribute (race, gender, nationality, etc.) that is the basis for defining the group. However, on attributes other than that central feature, the members of the category very likely will show enormous variation. Although all women, all blacks, all Italians are alike in one (group-defining) respect, these groups are huge, and their members are enormously diverse in virtually all other respects. We believe it is this fact that led our subjects, when asked to rate “how similar are the members of the group to each other,” to rate their similarity at a rather modest level. That is, despite colloquial sayings that imply broad similarity among social category members, people recognize the variability of such groups when explicitly rating their degree of similarity, especially in the context of rating the similarity among members of intimacy and task-oriented groups. Of course, and thirdly, the fact that social categories are not perceived as highly entitative or their members as highly similar has not prevented people— both our subjects and people more generally—from developing rich, well articulated stereotypes about many of these social categories. What we have, then, is a situation in which (a) people develop very important stereotypic conceptions (b) about social categories (but seemingly not about other types of groups?), (c) those social categories are not perceived as unified groups to the same extent as many other groups, and (d) those stereotypes are about groups whose members share some feature but otherwise may have little in common. In our view this convergence of properties poses some interesting dilemmas. First, it seriously questions the seemingly simple and comfortable alliances among the concepts of homogeneity, essentialism, and entitativity to which we alluded earlier. Second, it raises some interesting new questions about the factors that lead perceivers to form stereotypic beliefs about groups. In both our literature and in common parlance, the term stereotype typically refers to beliefs about a social category. We rarely use that term in talking about other types of groups (intimacy and task groups) that often are more entitative and homogeneous. Does this mean that we do not develop beliefs about those other types of groups? If we do possess such beliefs, what (if anything) is different about those belief systems when applied to social categories, such that the term stereotype is commonly used in such cases, but not others? In the following sections we discuss these questions. In some cases we can offer empirical evidence that speaks, at least partially, to the question under consideration. In other cases we can only engage in thoughtful analysis and offer our own perspective. What Types of Groups Get Stereotyped? The first question we raise stems from the observation that, when we think about stereotypes, we most commonly think about the groups representing Lickel et al.’s (2000) social category type. These are the groups about whom people develop elaborate stereotypes that often endow members of such groups with extensive lists of assumed attributes. And as we know from the rich literature of the last 20 years, these stereotypes function as cognitive structures that influence and guide information processing, judgment processes, and interpersonal behavior. Yet as we have noted, social categories are not perceived as being nearly as entitative as are many other groups, specifically, those that Lickel et al. have termed intimacy
38
THE PSYCHOLOGY OF GROUP PERCEPTION
groups and task groups. One might then wonder why, if intimacy and task groups are perceived as unified coherent units to such a greater extent, we do not have stereotypes of such groups. Or do we? To address these questions, we need to consider what we mean by the term stereotype. The literature contains many definitions of stereotypes, but a useful definition for our purposes is that a stereotype is “a cognitive structure that contains the perceiver’s knowledge, beliefs, and expectancies about a human group” (Hamilton & Trolier, 1986, p. 133). The issue here is, Do we have such cognitive structures only about groups that are social categories? Do we not also have “knowledge, beliefs, and expectancies” about other groups (families, juries, orchestras, labor unions, etc.)? It certainly seems plausible that perceivers have such cognitive structures, and if they do, then the issue becomes whether, and how, those structures differ from those we typically refer to as stereotypes. To explore these questions, we (Rodgers, Hamilton, & Sherman, 2003) set out to assess the stereotypes of a number of groups in each of two group types, namely, task groups and social categories. The first step was to assess the content of people’s beliefs about a variety of groups, using an open-ended methodology that permits participants to report spontaneously the content of the beliefs that they hold. This strategy was used because little is known about people’s perceptions of and expectancies about task groups, thereby requiring this initial stage for determining the content of “stereotypic” beliefs about such groups. The content and valence of people’s beliefs about 12 social categories (e.g., African Americans, Jews, women, etc.) and 12 task groups (e.g., an environmental organization, a sports team, a state legislature, etc.) were assessed using a free-description stereotype content methodology (Esses, Haddock, & Zanna, 1993). Prior research indicates that people’s knowledge, beliefs, and expectancies about social groups may be organized within a hierarchical system or taxonomy (Devine & Baker, 1991; Brewer, Dull, & Lui, 1981). Therefore, social categories and task groups of differing levels of specificity were included in the study (e.g., for social categories, a religious group, Jews, and Orthodox Jews; for task groups, a public interest organization, an environmental organization, and the Sierra Club). The participants were asked to think about each group and to write down any thoughts that came to mind. They were further instructed to list as many descriptors or short phrases that they believed were necessary to adequately convey their impressions of each group. In addition, participants assigned a valence rating to each of the attributes that they had listed. The extensive list of descriptors (more than 9,000) generated in this phase were reduced by coders who grouped descriptors into equivalent content categories. The descriptors generated by the participants were coded into 23 response categories, which included personality traits, physical attributes, demographic characteristics, typical behaviors, social roles, and goals/motives of the group. Traditionally, stereotype content studies have emphasized personality traits to the neglect of other content components (Ashmore & Del Boca, 1981). More recently, however, researchers have recognized the multicomponential nature of stereotypes and have argued that multicomponent assessment techniques are needed to more fully delineate the content of these cognitive structures. For example, some research indicates that behavioral tendencies and physical attributes are central to people’s stereotypic conceptions of certain groups (e.g., Deaux & Lewis, 1984). Therefore, the coding scheme used in this study included multiple aspects of content generated in participants’ free descriptions. Analyses of these data indicated that there were some significant differences in the nature and content of people’s cognitive representations of social categories and task groups. For example, participants ascribed significantly more descriptors to social categories than to task groups, and their descriptions of social categories included more references to contextual objects (e.g., “briefcases” and “business suits” for career women). This may suggest that people’s conceptions of social categories are more fully developed or more varied in content than are their conceptions of task groups. The attributes ascribed to social categories were
PERCEIVING THE GROUPNESS OF GROUPS
39
generally positive, but they were significantly less positive than those ascribed to task groups. Also, the range of the valence ratings was significantly greater among social categories than among task groups. On the other hand, and more importantly, the results also indicated that people’s cognitive representations of social categories and task groups were structurally quite similar in certain ways. For example, participants cited personality traits and affective reactions to roughly comparable degrees in characterizing task groups and social categories. Analyses of linguistic properties of these descriptions indicated that participants used adjectives more than nouns, and nouns more than verbs, when describing both social categories and task groups. Overall, personality traits and typical behaviors were listed most frequently by participants for both task groups and social categories. The fact that they ascribed enduring, stable qualities to both types of groups suggests that they have somewhat comparable cognitive representations for both group types. We also examined the degree of consensus among participants in their conceptions of the various groups by determining the percentage of respondents who generated the five most frequently cited descriptors for each group. Importantly, the consensus ratings for social categories and task groups were also very similar. The consensus ratings for these descriptors for social categories ranged from 16% to 40% (mean=26%), whereas the consensus ratings for task groups ranged from 15% to 50% (mean=28%). These consensus percentages are similar to those reported in other studies using a free-response methodology (Esses et al., 1993; Marin , 1984; Nieman, Jennings, Rozelle, Baxter, & Sullivan, 1994). The comparable values for social categories and task groups suggest that there is a comparable degree of consensus in people’s freeresponse descriptions of social categories and task groups. Taken together, the results of this study suggest that individuals hold shared beliefs (or “stereotypes”) about various task groups and that these beliefs are structurally similar to those held about traditional social categories. Thus, people have generalized beliefs and expectancies (stereotypes) about task groups as well as social categories. It seems likely that people have similar belief systems and expectancies about intimacy groups as well. If so, then the conclusion would be that perceivers have generalized cognitive structures pertaining to each of the different group types identified in Lickel at al.’s (2000) research. If that is the case, then some interesting new questions arise. Entitativity, Essentialism, and Stereotypes One important question that immediately arises is, What is special about social categories such that we refer to belief systems about them as stereotypes, and how might those generalized expectancies differ from the beliefs held about other types of groups? One possibility, alluded to earlier, is that perceivers view (some) social categories as possessing some “essence,” some inherent qualities, perhaps even reflecting some biological basis of group membership. And if that is the case, then one might assume a close relation between the perception of essentialistic bases of group membership and the perception of group entitativity. In line with this thinking, Yzerbyt, Rocher, and Schadron (1997) proposed that, when perceivers are faced with a highly entitative group, they will attempt to extract the underlying essence of that group. Thus, perceptions of entitativity of a group enhance essentialistic thinking about the group, which in turn will affect perceptions of individual members of the group through a stereotyping and generalization process. As a consequence, highly entitative groups are associated with stronger dispositional inferences (Yzerbyt, Rogier, & Fiske, 1998) and greater correspondence bias (Rogier & Yzerbyt, 1999) than are low entitativity groups. Despite the fact that social categories are perceived in the low to moderate range on both entitativity and homogeneity, this group type seems to be quite high on the dimension of essentiality. In their discussion of
40
THE PSYCHOLOGY OF GROUP PERCEPTION
natural kinds, Rothbart and Taylor (1992) argue that certain social groups are perceived as natural kinds, and natural kinds are characterized by high unalterability and high inductive potential. The types of groups that Rothbart and Taylor view as natural kinds are social categories that have a biological basis. This would include social categories such as race and gender. Rothbart and Taylor discuss the implications for essentiality and stereotyping of viewing these social categories as natural kinds. It is especially interesting that social categories such as race and gender are viewed as natural kinds with a distinct essence in light of the fact that these groups are not seen as particularly unified or homogeneous. Is it simply the biological basis of these social categories that leads to a perception of essentiality? And is this factor sufficient to understand why we develop richly articulated stereotypes about social categories? Although some instances provide compelling examples for this argument, particularly categories based on gender and race, what about other social categories about which perceivers develop and use stereotypes? Is there some underlying (biological) essence that defines social categories based on religion (is there an essence shared by Catholics?), nationality (is there an essentialism associated with being American?), or occupation (is there some inherent quality associated with truck drivers?)? Do perceivers endow these groups with essentialistic qualities, and is that what underlies stereotypic views of them? These latter examples suggest that, although some plausibility clearly exists, there may be constraints on the extent to which essentialism can be a useful basis for understanding stereotypes. To carry our analysis a step further, let us consider another type of group, the intimacy group. A family is a prime example of an intimacy group. And in fact family members are generally seen as very high in both entitativity and similarity. In addition, families certainly have a far greater biological link (the sharing of a significant proportion of the gene pool) than do broad social categories. Yet it seems to us intuitively that there is less stereotyping of a group of family members and less generalization from one family member to another than there is for broad social categories. The question of “What are women like?” or “What are Italians like?” seems more sensible and more open to an answer based on a variety of traits and attributes than does the question “What is the McMillan family like?” Thus, even in cases where some biological essence can plausibly be assumed, the link to stereotypes is tenuous. In sum, notions of essentialism may play an important role in at least some instances of stereotyping. However, the frequency of (a) cases in which stereotypes are common but it is difficult to assume essentialism and (b) cases in which biological essence can easily be assumed, but stereotyping is not evident suggest a need for caution in applying such conceptualizations. These distinctions among essentialism, entitativity, and stereotyping again support the view that we are dealing with quite different and at least somewhat independent concepts. The Nonequivalence of Entitativity and Homogeneity Many of the researchers who have generated the recent renewal of interest in the concept of entitativity have come to this topic from a background in stereotype research. This is not surprising, in that entitativity has to do with perceptions of a group, and within social psychology the study of stereotypes has for many years been the primary focus of research concerning the perception of groups. It seems natural, then, that these scholars would find an easy link between the perception of group entitativity and the perception of group homogeneity. Indeed, a reviewer commenting on one of our earlier papers made the following observation: “Why do we need the construct of entitativity? I have never really understood what makes entitativity different from group homogeneity.” This near-equivalent relation is implicit in some research on entitativity by several researchers. To illustrate, consider some recent empirical studies. For example, participants in Brewer, Weber, and Carini’s
PERCEIVING THE GROUPNESS OF GROUPS
41
(1995) research saw a videotaped discussion among six persons, three of whom were shown wearing red sweatshirts while the other three were wearing blue sweatshirts. In some conditions the discussion groups were differentiated only in terms of the color of their shirts, whereas in other cases these groups were also said to reflect personality distinctions or different fields of study. Thus, group entitativity was based on one or more dimensions on which group members were similar. After viewing the videotape, the participants’ task was to identify which person had made which comments during the discussion, and recognition errors were coded as to whether they were within- or between-group errors (S.E.Taylor, Fiske, Etcoff, & Ruderman, 1978). Overall, participants made more within-group than between-group errors, indicating that the group distinction was used in processing and storing the information. However, when similarity of sweatshirt color was the only basis presented for distinguishing between groups, this difference did not emerge. Thus, in this research anyway, the manipulation of entitativity in terms of visual similarity by itself (in the absence of other bases for group membership) did not produce the intergroup differentiation outcome, but when combined with other bases of similarity, it did produce effects. In another study reflecting this assumed equivalence, Yzerbyt, Rogier, and Fiske (1998) manipulated group entitativity by describing sets of three persons as students either at the same college or at different colleges. Thus, the only difference between the group and aggregate conditions was whether the three individuals were similar on one feature, that being the college they attend. Participants observed two such groups perform the questioner-answerer quiz game used by L.D.Ross, Amabile, and Steinmetz (1977). Results showed that perceptions of the general intelligence of the questioners and answerers differed most strongly when the questioners were identified as being students at the same school and the answerers were all students at a different school. Thus, perceived entitativity, defined solely in terms of a single shared feature, had significant effects on the attribution process. More recently, Dasgupta, Banaji, and Abelson (1999) manipulated the physical similarity among stimuli in order to determine the effects of group entitativity on judgments of the group. In one experiment, entitativity was varied in terms of similarity of the skin color of faces (variability within the group). Using a color scale, participants then estimated the average and the variability of the skin color of the faces seen. Results showed that participants were more accurate in judging color for the entitative than the nonentitative group. In a second study, participants were shown groups of humanoid, science-fiction-like figures that varied in color, thus again manipulating entitativity in terms of similarity among group members. The task of participants was to form an impression of the group and to judge it on several personality attribute scales. Participants made significantly more negative judgments about the more entitative group. All of these studies are very fine examples of investigations of the effects of similarity manipulations on various dependent measures. However, it is important to note that none of these studies included any manipulation checks or other direct measures of perceived group entitativity. Therefore, the interpretation that perceived group entitativity produced the reported effects on the various dependent measures rests on the assumption that the similarity manipulation had its influence on these outcome variables by inducing differing levels of perceived group entitativity. Our purpose in describing these studies is not to suggest that the authors equate similarity with entitativity at a conceptual level; this clearly is not the case (cf. Abelson et al., 1998; Brewer & Harasty, 1996; Yzerbyt, Rocher, & Shadron, 1997). Rather, our point is that researchers need to include measures to verify that the perception of group entitativity is in fact the influential factor in understanding obtained outcomes. Thus, for example, Castano, Yzerbyt, & Bourguignon (in press) have reported evidence of the mediating role of entitativity in the effects of similarity on group judgments. More importantly, although similarity among group members may indeed foster the perception of entitativity in the group, and although the perception of group entitativity may increase perceptions of
42
THE PSYCHOLOGY OF GROUP PERCEPTION
similarity, these facts alone do not justify equating similarity with entitativity. There are several important reasons for avoiding this pitfall. The first basis for this caution goes back to D.T.Campbell’s (1958) analysis in which similarity among individual members was one, but only one, basis on which entitativity could be perceived. Campbell also identified several other properties that could serve as cues to entitativity (interaction among individuals, common goals, common fate, etc.), most of which do not require similarity among those members. Indeed, some authors (Hamilton et al., 1998) have argued that similarity is not the most important of those properties. In fact, Lickel et al. (2000) empirically showed that several of these variables are equally correlated with ratings of group entitativity across a broad spectrum of groups. Also, L.Gaertner and Schopler (1998) have shown that interaction among group members has a strong effect on perceived group entitativity. Thus, similarity may be an important cue to entitativity, but only in some cases and under some conditions. A second basis for concern comes from Lickel et al.’s (2000) findings of different types of groups that differ in perceived entitativity. Intuitively, it seems that similarity would be an important property that implies a high degree of “groupness” for groups that represent social categories. However, for other group types (intimacy groups, task groups) similarity among group members may not be an important groupdefining property. In fact, early scholars in the group dynamics tradition (Kretch & Crutchfield, 1948; Lewin, 1948) argued strongly that similarity is not a basis for defining groups, but rather interdependence (reflected in common goals and outcomes) is the most important feature of real groups. Thus, homogeneity may be a more important basis for group perception for social categories than for other group types (including more entitative groups). Third, there are accumulating empirical findings that argue persuasively that the perception of entitativity may, under certain conditions, be rather independent of group homogeneity. The results of two studies by Welbourne (1999) support this view. In both studies participants read a series of sentences describing behaviors performed by members of a group, and they subsequently made judgments of the group. Prior to reading the stimulus sentences, group entitativity was manipulated through paragraphs describing the group. In one experiment, these paragraphs described the members of the group as being high or low in similarity and behavioral consistency. This manipulation did not have any significant effect on participants’ judgments of the groups. However, in another experiment, entitativity was manipulated by describing the group as high or low in the extent to which it is organized and as having a common goal and purpose. In this case the entitativity manipulation had a significant effect on group impressions. Thus, in the same paradigm, entitativity operationalized as similarity and consistency had no effect, whereas it did when operationalized in terms of organization and coherence (see also Hamilton et al., 1998). Other studies have compared high and low group entitativity conditions in participants’ judgments about those groups and have specifically sought to investigate the mediating processes underlying such effects. For example, Thakkar (2003) manipulated entitativity through introductory paragraphs that described the group either as tightly structured, with high degrees of interaction as members pursued common goals, or as loosely structured, with infrequent interactions as members worked independently in completing the group’s objectives. Participants then read behavior-descriptive sentences describing group members, their task being to form an impression of the group. For half of the participants the information was predominantly favorable; for the other half, predominantly unfavorable. Thakkar reasoned that, if entitativity induces more integrative processing (Hamilton et al., 1999), then participants should make stronger inferences from the behavioral information—and hence make more extreme judgments of the group —in the high compared to the low entitativity condition. After reading the behavior-descriptive stimulus items, participants rated the group on several attributes reflected in the information provided. Thakkar’s
PERCEIVING THE GROUPNESS OF GROUPS
43
findings supported this hypothesis: Participants made more extreme judgments (positive or negative) in the high entitativity condition than in the low entitativity condition. Thus, perceived entitativity leads to more polarized impressions of the group. What accounts for this polarization effect? To evaluate this question, Thakkar conducted mediational analyses. Specifically, participants rated the group not only on relevant attribute dimensions but also on a variety of measures of group entitativity and related properties. Group entitativity was assessed by the mean of ratings on four scales: how unified the group is, how cohesive the group is, how much the members feel a part of the group, and the extent to which the group “qualifies as a group.” Perceived group similarity was assessed by a rating of how similar or dissimilar the group members are to each other. Mediational analyses showed that the influence of manipulated entitativity on judgments was significantly mediated by perceptions of group entitativity (as assessed by these measures), but not by participants’ ratings of similarity among group members. This divergence in mediational measures suggests caution in equating entitativity with perceived homogeneity. Thakkar’s (2003) research demonstrates that entitativity leads to stronger, more polarized judgments of a group. Does entitativity also increase the generalization of inferences from one group member to another? To investigate the role of perceived entitativity on the generalization of attributes among group members, Crawford, Sherman, and Hamilton (2002) employed a savings-in-relearning paradigm (Carlston & Skowronski, 1994). Participants viewed photographs and read behaviors performed by members of two groups. For each group, half the members were paired with behaviors that implied one specific trait (e.g., aggressiveness), and the other half engaged in behaviors that implied a different trait (e.g., honesty). The two groups were both described as either high or low in entitativity. During this initial phase, participants would have the opportunity to form a general conception of each group. In the next phase, the photographs were paired either with a trait term that matched the behavior that was originally linked with that target (a trait inference trial), or with a trait term that matched the behaviors of other members of the target’s group (a trait transference trial). Results showed that inference trials, but not transference trials, were learned well for low entitativity groups, whereas transference trials were learned far better for high entitativity groups. In fact, participants performed equally well on inference and transference trials for high entitativity groups. These results support the idea that, when perceived entitativity is high, participants extract a general impression of the group and view the individual members as interchangeable parts. The members lose their individuality, and the traits of any member are transferred to all other members. For low entitativity groups, on the other hand, members are treated as unique individuals, and information related to specific individuals is stored separately in memory. Earlier we summarized a study (Rodgers et al., 2003) showing that perceivers have rather similar kinds of cognitive representations—knowledge, beliefs, and expectancies—for task groups and social categories. In this sense, we can consider that perceivers have “stereotypes” of task groups, comparable in many respects to those they have for social categories. In a subsequent study, we examined the relations among perceptions of entitativity, homogeneity/similarity, essentialism, and stereotyping for both task groups and social categories. Participants rated four social categories (e.g., African Americans, Jews) and four task groups (e.g., a public interest organization, doctors) at two levels of specificity (e.g., a public interest organization, Sierra Club). The free-response descriptors generated by participants (summarized earlier) were used to create rating scales on which participants rated the groups. For each group, we identified those descriptors that had been most commonly ascribed to that group, and a participants ratings on those attributes, determined uniquely for each group, were averaged to form a group-specific stereotype score for that group in the previous study. Also for each group, a participant’s ratings on the remaining descriptors were averaged as a non-group-specific stereotypic score. Perceptions of entitativity were assessed by eight
44
THE PSYCHOLOGY OF GROUP PERCEPTION
items tapping perceptions of interaction, group importance, unity, organization, interdependence, cohesive ness, members feeling part of the group, and the extent to which the group “qualifies as a group.” Three scales reflecting similarity in appearance, personality attributes, and behaviors were averaged to assess perceived homogeneity. To measure essentialism we used 7 items adapted from Haslam et al. (2000). In addition, we measured perceptions of a group’s role differentiation and agency (four items assessing each construct). Overall, perceptions of entitativity were higher for task groups than for social categories, replicating Lickel et al.’s earlier results. In contrast, both perceived homogeneity and essentialism were significantly higher for social categories than for task groups. For both role differentiation and agency, task groups were rated significantly higher than social categories. How well do these variables function as cues to group entitativity? All four of them were significant predictors of perceived entitativity for both task groups and social categories, although these coefficients were generally higher for role differentiation and agency than for homogeneity and essence. We then determined the extent to which these four predictor variables (homogeneity, essence, role differentiation, agency) were correlated with group stereotyping, as assessed by ratings on the groupstereotypic descriptors discussed earlier. In particular, we conducted a series of mediational analyses to determine the extent to which perceived entitativity mediated the relationship between each of these variables and stereotyping. These analyses were conducted separately for social categories and task groups. The overall pattern of results can be summarized as follows. For social categories, all four predictor variables were significantly correlated with the strength of participants’ ratings on the attribute dimensions, and in all cases those relationships were substantially reduced when the mediating role of entitativity was controlled. For task groups, role differentiation and agency again were significant predictors of stereotyping, and those relationships were reduced when entitativity was controlled. In contrast, homogeneity did not predict stereotype ratings. The extent to which essence predicted stereotyping varied across replications, but in contrast to other variables, entitativity did not mediate the relationship between essence and stereotyping for task groups. Thus, the concepts of entitativity, homogeneity, and essentialism play somewhat divergent, though related, roles in guiding people’s perceptions of groups. These constructs are correlated with each other, but the Rodgers et al. (2003) findings imply that homogeneity and essence are more strongly related to perceptions of social categories, whereas role differentiation and agency are more directly pertinent to judgments of task groups. Importantly, perceptions of group entitativity played an important mediating role in the extent to which any of these variables was related to judgments of these groups on a variety of groupdescriptive attributes. In general, these results indicate not only that entitativity, homogeneity, and essentialism are distinct concepts but also that they may have different relations and functions in group perception for different types of groups. CONCLUSIONS In this chapter we have focused on several concepts that have been used almost interchangeably in recent analyses of group perception: entitativity, homogeneity, and essentialism. Some authors have treated them as near-synonyms; others have maintained some conceptual distinctions but have emphasized close relations among them, viewing one concept as a close antecedent of or causal factor for another concept. Our view, in contrast, is that, although these constructs may be closely related in some contexts, may be correlated to some extent, and may interact with each other in influencing group perceptions, they are certainly not
PERCEIVING THE GROUPNESS OF GROUPS
45
identical. In fact, we would argue that they are relatively independent concepts and refer to different aspects of group perception. Why do we argue that they are distinct concepts? Several types of evidence—some empirical, some conceptual—support this conclusion. Earlier we reported findings that strongly suggest that entitativity and homogeneity can play different roles in guiding and mediating group perceptions. In addition, it is instructive to consider the implications of several other findings. First, a group could be high in entitativity but only low or moderate in perceived similarity. This pattern would be true for many task groups that have a clear division of labor and role differentiation. The members of such groups can be quite varied in their talents, traits, and other attributes, yet the group as a whole can be highly entitative. Such groups would also be low in essentialism. Second, a group could have only modest levels of entitativity and similarity but be high in essentialism. This pattern could easily be true for some social categories. Third, there may be groups (e.g., some social categories) that are perceived as being quite high in similarity and thus are homogeneous, but are not seen as possessing an underlying essence and are perceived as only moderate in entitativity. All of these (and other) possibilities indicate the need to maintain distinctions among the concepts. A second question of obvious importance concerns the relationship of each of these concepts, and particularly their differential relationships, to stereotyping. We have alluded to this issue at several points, but we believe these relations are, at this point, far from clearly established. Perhaps the simplest element in this conceptual equation is the relationship between perceived homogeneity and stereotyping. The perception of homogeneity in a group facilitates making generalizations about the group and its members, and ever since G.Allport’s (1954) classic analysis, overgeneralization has been recognized as one of the primary characteristics of stereotyping. But what of the other relations? Certainly recent authors (Rothbart & Taylor, 1992; Yzerbyt, Rocher, & Schadron, 1997) have argued that (at least some) stereotyped groups are perceived as possessing some underlying essence that affords greater inductive potential. Does it also increase perceptions of group homogeneity and the generalization of attributes across group members? We need more evidence to satisfactorily ad dress these questions. Finally, what is the relationship between entitativity and stereotyping? This question also awaits the relevant evidence, but it is worth noting again that social categories—the most stereotyped of the group types uncovered in Lickel et al.’s (2000) analyses —are perceived as being only moderate in entitativity. Thus, a number of crucial questions remain about the relations among these constructs. Yet it is clear that all of them are at the heart of group perception. The extent to which a collection of individual persons is perceived as constituting a viable group is fundamental to any and all subsequent phenomena of group perception. The extent to which the group is seen to be homogeneous; the extent to which members of the group are perceived as possessing other characteristics that comprise a group stereotype; the possibility that the group perception includes some inherent essence that defines what the group represents—all of these phenomena require and build upon the perception that a set of people constitute a group. In that sense the perception of group entitativity— including the nature of the ties that bind people into the group, the type of group it is, and its function and purpose—seems to us to be a fundamental element in group perception, a precursor to many of the other aspects of group perception that are of both theoretical and practical interest. We hope that our analysis in this chapter has highlighted the need to maintain distinctions among these concepts. We also emphasize the need for researchers to be clear in what they mean by these terms as they use them, to develop means of manipulating and implementing these constructs in experimental procedures, and to include measures of these concepts in their experiments. Through these procedures we can, in a relatively short time, rigorously gather the data needed (hopefully converging from different labs) that will clarify the meanings of and relations among these important constructs.
46
THE PSYCHOLOGY OF GROUP PERCEPTION
AUTHOR NOTE Preparation of this chapter was supported in part by National Institute of Mental Health Grant MH-40058 to the first two authors and by Grant K05DA00492 to the second author.
4 Essentialism and Entitativity Structures of Beliefs about the Ontology of Social Categories NICK HASLAM University of Melbourne LOUIS ROTHSCHILD Brown University DONALD ERNST Hillsdale College
“Oh dear,” Gloria said. “You’re not a nominalist…. You’re a silly old essentialist!” —Cathleen Schine, The Evolution of Jane (1998, p. 130) “Essentialism” rarely arises in polite conversation, even between characters in novels, and the chances of “entitativity” entering colloquial English seem slim indeed. Nevertheless, as the pages of this volume attest, the two terms have become indispensable for psychologists and other social scientists who study group perception. There are many reasons behind the growing interest in the two concepts, but the rich theoretical possibilities that they offer, and the connections to other disciplines that they open up, are surely among the most important. What makes these concepts so interesting to work with, particularly in the case of essentialism, is the sheer range of theoretical contexts in which they have been employed and the complexity of meanings that they have acquired in the process. In this chapter, we will review the main strands of work on essentialism and entitativity, propose an integrative framework for them, and describe some empirical research based on it. After discussing the implications of our findings for the structure of fundamental beliefs about social categories, we will close with some thoughts on future directions for the study of the two concepts. ESSENTIALISM The term essentialism appears to have been coined by the philosopher Karl Popper in the 1930s (see Popper, 1982), although the notion of “essence” has a history in philosophical discourse that stretches to Aristotle and beyond. For Popper, essentialism characterized any intellectual position that took concepts to have precise, ascertainable meanings, and the associated methodological stance that takes attempts to
48
THE PSYCHOLOGY OF GROUP PERCEPTION
explicate, fix, and sharpen word meanings to be important steps toward reliable knowledge. Popper vigorously opposed these positions, espousing an antiessentialism in which word meanings are considered fundamentally inexact and ambiguous, prone to change over time, and a function of the context in which the word is used. Consequently, in the quest for knowledge we should avoid quarreling about word meanings and focus instead on determining the truth of propositions and theories. Although Popper’s original analysis of essentialism differs in several respects from later uses of the term, and has not been widely taken up since, it provides a good starting position from which to survey essentialism’s many later developments, and shares several elements with them. First, his understanding of essentialism relates it to the nature of concepts and in particular to the question of whether concepts can be anchored by defining features. Second, he presents essentialist positions as those that imply that word meanings have sharp boundaries based on fixed and context-insensitive criteria. Third, he takes a resolutely critical approach to such positions, arguing for the ambiguity, mutability, and undefinability of word meanings, for example, anticipating the reluctance of all but a few later thinkers to identify themselves as essentialists. These later thinkers can be sorted into a few distinct and often intellectually isolated traditions, each of which developed distinct but overlapping understandings of essentialism. These traditions will be briefly reviewed in turn. For a more extensive and critical review see N.Haslam (1998). Philosophy of Language Popper’s sense of essentialism most closely resembles the one developed as part of the critique of the socalled classical or Aristotelian account of concepts. By this account, each concept has a set of necessary and sufficient, or defining, properties. These essential properties make the things referred to by the concept the kinds of things that they are. Concepts only apply when all of their essential features are present, so that concepts refer in an all-or-nothing fashion to a discrete and sharply bounded class of phenomena. The antiessentialist critique of this account of conceptual structure derived largely from the work of Wittgenstein (see Hallett, 1991), who presented word meanings as based on family resemblances among category members rather than on precise definitions. By this account, concepts need not have a set of defining properties, and may apply by degrees rather than in any absolute way, so that category boundaries are imprecise and ambiguous. Eleanor Rosch (e.g., 1975) famously introduced this view of concepts into psychology in the form of prototype theory and showed how the classical account of categorization failed to account adequately for a variety of conceptual phenomena. Prototype theory was put to service in an influential critique of essentialism and the objectivist view of reality associated with it by the linguist George Lakoff (1987). There has been much debate since among cognitive scientists on the pros and cons of the classical and prototype accounts of concepts. However, the important issue for our purposes is that in this linguistic tradition of work, essentialism is a matter of word meanings, and in particular the existence of defining properties that give rise to sharp conceptual boundaries. Moreover, there is evidence (Medin & Ortony, 1989) that people’s thinking about word meanings is in this fashion essentialist, whether or not this represents a linguistic or metaphysical error. The Philosophy of “Natural Kinds” A related but importantly distinct sense of essentialism springs from the work of a number of thinkers who propounded a new theory of direct reference and a “naturalized epistemology.” Philosophers Saul Kripke (1980) and Hilary Putnam (1975) developed a position that some categories, dubbed “natural kinds” and often exemplified by biological species, chemical elements, or compounds, such as gold, water, or tigers, do
ESSENTIALISM AND ENTITATIVITY
49
indeed have essences. Moreover, the meaning and applicability of the corresponding concepts is appropriately fixed by these scientifically describable essences, rather than by the superficial, observable features that laypeople’s concepts may employ. Proponents of the concept of natural kind such as Hilary Kornblith (1993) argue that humans may have evolved to think about natural entities in an essentialist fashion because this accurately captures the underlying nature of these entities. They also maintain that because natural kinds share deep similarities, “carving nature at the joints,” they are particularly rich sources of inferences, having what philosophers call “inductive potential.” Although the essentialism defended by these philosophers of natural kinds appears on the surface to be much the same as the one implicit in the classical view of concepts, there are some important differences. For one, it is restricted to the domain of natural entities, rather than to word meanings in general. For another, the essence of a natural kind is not a matter of word meanings per se, but of some sort of real underlying structure—genetic, atomic, or molecular— that gives rise to the kind’s outward properties and can in principle be determined scientifically. Consequently, as Putnam (1975) argued, “what the essential nature [of a thing] is is not a matter of language analysis but of scientific theory construction” (p. 104). The concept of natural kind, and its associated sense of essentialism, has also found its way out of philosophy and into contemporary developmental psychology and anthropology. Considerable research has explored the ways in which people reason differently in the conceptual domains of “living kinds” and of human artifacts (e.g., Keil, 1989), and studies have frequently found that laypeople in many societies (Atran, 1990) hold essentialist theories about living kinds but not artifacts. This ontological distinction is made even by very young children (e.g., Gelman, 1988). For instance, young children recognize that an animal still belongs to its kind despite changes to its appearance that make it resemble another kind, but they see equivalent transformations to an artifact as changing its identity. Philosophy of Biology In the philosophy of biology, essentialism has been used to refer to the pre-Darwinian species concept, according to which each species has an eternal, changeless nature shared by all specimens of its kind (Dupré, 1981; Hull, 1965; Mayr, 1982). This view, which is substantially similar to the folk-biological theories of the world’s traditional societies (Atran, 1990), is contrasted with the now dominant scientific view, in which species are historically changing, internally varying populations, whose members may share no necessary properties and may grade into other species under certain conditions. Critical Social Theory Essentialism has become a particularly popular term in recent social science and cultural studies, especially in connection with the critique of theories of gender, race, and sexual orientation (e.g., Fuss, 1989; Grosz, 1990; Stein, 1990). Theories or intellectual positions are typically branded essentialist when they claim that these social distinctions have deeply rooted biological underpinnings, that they are historically invariant and culturally universal, or that their boundaries are sharp and not amenable to sociocultural shaping. This critique of essentialist positions is often carried out in the name of social constructionism. The antiessentialist critique consequently emphasizes, often through historical, sociological, and anthropological scholarship, the ways in which our classifications of one another are cultural artifacts, ways of dividing up the world that are contingent on language, history, and human interests. Moreover, antiessentialist writers typically claim that essentialist beliefs are linked to stigma and prejudice because they make the
50
THE PSYCHOLOGY OF GROUP PERCEPTION
subordination of some kinds of people seem inevitable or naturally ordained. Essentialist beliefs, that is, are seen to have an ideological or system-justifying function. It must also be noted, however, that some workers in the tradition of critical theory take a more positive stance to ward essentialism, including some French feminist writers and a strain of gay activists who promote a biologically grounded “strategic” essentialism as a way of forging a strong political identity and disarming those who present homosexuality as a freely chosen perversion. The theoretical traditions reviewed above do not exhaust the intellectual contexts in which the concept of essentialism has been deployed. For instance, Anglophone intellectuals often neglect the ways in which the French existentialists were explicitly antiessentialist with respect to social categories. Sartre opposed existentialism to essentialism, declared humans the beings whose essence it is to have no essence, and analyzed the pathologies that result when people treat themselves or others as essentialized “in themselves” (en-soi) rather than self-determining beings “for themselves” (pour-soi). Nevertheless, traditions presented here cover the primary domains of work on essentialism that have influenced recent work on social categorization. ENTITATIVITY The literatures on entitativity are fewer and less multidisciplinary than those on essentialism, although they arose out of quite varied inquiries about the nature of groups and of the “group mind.” The term was coined by a psychologist (D. T.Campbell, 1958) and adopted without extensive reinterpretation by his psychological successors (e.g., Abelson, Dasgupta, Park, & Banaji, 1998; Brewer & Harasty, 1996; Hamilton & Sherman, 1996; McGarty, Haslam, Hutchinson, & Grace, 1995). By “entitativity” Campbell meant the perception that a social aggregate is a coherent, unified and meaningful entity, a perception grounded, he argued, in Gestalt principles of similarity, proximity, pregnanz, and common fate. Later writers have contributed to the characterization of the phenomenon, showing, for instance, that entitativity judgments are associated with perceived group homogeneity (McGarty et al., 1995) and behavioral consistency (Hamilton & Sherman, 1996) and differ for different types of group (Lickel et al., 2000). Entitative groups are understood to have a coherence that approaches that of individuals (McConnell, Sherman & Hamilton, 1997); and entitative groups promote the drawing of dispositional inferences about group members (Yzerbyt, Rogier & Fiske, 1998). Although the concept of entitativity appears to have a rather straightforward pedigree, it is not without the sorts of precedents and parallel conceptual developments that we observed in the case of essentialism. Indeed, a concern with the ways in which people understand the social world to be composed of real entities has been a longstanding focus of constructionist sociologists, as far back as P.L.Berger and Luckmann’s The Social Construction of Reality (1966/ 1971). Within this tradition, “reification” performs much the same conceptual function as Campbell’s entitativity, referring to the granting of some sort of concrete or thinglike reality to a social category or type, the attribution to members of the category of an intrinsic sameness and an objective status independent of particular perceivers. Here are Berger and Luckmann (1966/1971) on the reification of social identity: Identity itself may be reified…both one’s own and that of others. There is then a total identification of the individual with his socially assigned typifications…. [R]eifications bestow an ontological and total status on a typification that is humanly produced and that…objectifies but a segment of the self. (p. 108)
ESSENTIALISM AND ENTITATIVITY
51
Thomason (1982) offers a useful illustrative treatment of the concept of reification as it appeared in the work of the sociologist Alfred Schutz. Schutz wrote of “a kind of ‘thingification’ which operates at the level of everyday reality and commonsense conceptions of the world” (Thomason, 1982, p.27). These commonsense understandings involved perceptions of the social world on the basis of types (“typifications”), such that we ascribe a standardized sameness to exemplars of each type and take types to refer to a nonperceptual order of reality. Reification therefore referred, for Schutz, both to a psychological process and an ontological judgment about the nature of social categories. From Schutz’s phenomenological perspective, this conferring of an “autonomous externality” on the social world serves a vital function by giving people a taken-for-granted sense of a shared social reality. Despite their manifest similarities, there are two primary difference between the concepts of reification and entitativity. First, reification is not limited to social categories; it refers broadly to the attribution of thinglike objectivity or concreteness to all manner of categories and abstractions. Second, sociological writers often present decidedly critical analyses of social reification, in contrast to Campbell’s quite neutral discussion of entitativity. To Thomason (1982), for example, reification “refers to the way people deny, forget or ignore the constructedness of their social worlds” (p. i) and “grant an improper thing-like givenness to the world” (p. 1). Reification distorts and falsifies social experience by bestowing on social types an “inappropriate ontological fixedness” (p. 88), and represents a theoretical error for which constructionist sociologists take their realist colleagues to task. In short, reification is presented in this theoretical tradition as a problematic cognitive process both for laypeople and for social theorists, as much as it may be deeply rooted in human psychology. This brief discussion of reification suggests that entitativity is not quite as original a concept as it might seem and that Campbell’s famous paper represents just one, innovative attempt to grapple with issues that have been worked over by other disciplines, chiefly sociology and anthropology. Psychologists interested in social categorization might broaden the theoretical context of their empirical work on entitativity by acquainting themselves with the literature on reification. Doing so might convince us, for instance, that entitativity may involve not just a perception of “groupness” or coherence, but also a nonperceptual judgment about the ontological status or nature of the category in question. The purely perceptual understanding of entitativity is encouraged by Campbell’s Gestalt perspective and the connections of entitativity research to research on group homogeneity. However, if an ontological judgment is an important element of entitativity, then such an understanding, based only on perceived similarity or consistency, will be lacking, and we will need to take seriously the ways in which people attribute underlying, intrinsic, or inherent bases and properties to entitative categories. Similarly, if the literature on reification is any guide, we might do well to be more attuned to the ways in which entitativity judgments are problematic, have adverse social consequences, and are associated with distortions and biases. EMPIRICAL WORK We have attempted to show that essentialism and entitativity are complex concepts that have figured in some form or another in many different theoretical contexts and have picked up distinct meanings along the way. However, those of us who study these concepts in relation to group perception have seldom attempted to clarify the exact meanings of the terms we are employing in our work or paused to ask about the structure of people’s essentialist or entitative thinking about social categories. In our opinion, that is, social psychological investigations of essentialism, in particular, have proceeded without sufficient conceptual analysis of what essentialist thinking involves or how it is structured.
52
THE PSYCHOLOGY OF GROUP PERCEPTION
In view of the lack of systematic attention paid to determining the structure of peoples essentialist thinking about social categories, we embarked on a series of studies. Based on the wide-ranging review of literatures on essentialism that we have presented above, we isolated a set of conceptual elements that appeared to us to be aspects or facets of essentialism in at least one theoretical tradition. From the philosophy of language we extracted the idea that essentialism implies (1) the existence of necessary or defining characteristics, and (2) the existence of sharp boundaries around, and all-or-nothing membership of, essentialist categories (i.e., discreteness). From the philosophy of natural kinds, we extracted the idea that essentialist categories are (3) in some sense “natural,” (4) have great inductive potential (i.e., afford many inferences about the properties of category members), and (5) imply immutability, such that instances of the category cannot cease to be instances while they retain its essence. From the philosophy of biology and critical social theory we extracted the idea that essentialist categories are (6) historically invariant. (This differs from immutability in referring to changelessness in the category itself rather than its members, and in its reference to historical time.) Permeating all of these literatures, we would maintain, are three additional aspects of essentialism: essentialist categories are (7) homogeneous or uniform because all category members are funda mentally alike, (8) based on inhering, intrinsic, or underlying properties rather than merely superficial resemblances, and (9) identity determining, such that belonging to an essentialist category excludes its members from other categories. By casting a wide net in this manner, attempting to encompass all of the important elements of essentialism, we hoped to be able to determine empirically which elements are centrally involved in essentialist thinking about social categories and how these elements are empirically interrelated. These nine essence-related elements can serve as a framework for making sense of essentialist thinking and for integrating other theorists’ uses of essentialism. For instance, the sense of essence implicit in the “entity theory” perspective of Dweck and her colleagues (e.g., Dweck, 1999; Levy, Stroessner, & Dweck, 1998) focuses exclusively on the immutability element, although they focus on the immutability of category members’ attributes rather than category membership itself. Rothbart and Taylor’s (1992) influential theoretical treatment of essentialist thinking about social categories focuses on two elements: immutability and inductive potential. Hirschfeld (1996) emphasizes the elements of inherence, discreteness, and naturalness in his analysis of essentialist racial categorization, and Yzerbyt, Rocher, and Schadron (1997) list the elements of defining features, immutability, inductive potential, homogeneity or coherence, and exclusivity. Study 1: The Structure of Shared Beliefs about Social Categories Taking our nine elements of essentialist thinking as a starting point, we first set about asking the basic descriptive question of how these diverse essence-related beliefs are organized and interrelated (Haslam, Rothschild, & Ernst, 2000). Do essentialist beliefs form a single integrated grouping, or are they structured in a more complex and differentiated way? In addition, we asked whether these beliefs are held especially about stigmatized categories, as many a social theorist has argued. These questions have not been systematically addressed in previous work, which has often been based on theoretical argument and review (e.g., Rothbart & Taylor, 1992) or has limited empirical attention to racial or gender categories (e.g., Hirschfeld, 1996). We examined these questions in a simple study in which 40 American undergraduates rated a sample of social categories on a set of items corresponding to our nine elements of essentialist thinking, as well as one evaluative item about the social status of the categories (i.e., “how valued or favorably regarded they are in our culture”). Forty categories were rated, systematically sampled in pairs from 20 domains, including sex, age, race, ethnicity, class, occupation, sexual orientation, interests,
ESSENTIALISM AND ENTITATIVITY
53
personality, political belief, religion, psychiatric disorder, and medical disease. Each participant rated one member of each pair. As our focus at this point was on shared beliefs and differences between categories, mean ratings of the categories on the nine items, across the 20 participants who rated each category, were then intercorrelated. The results clearly indicated that essentialist beliefs do not represent a unitary set. Instead, the nine elements of essentialism formed two distinct sets, correlations within which were very strong (mean r=0. 64) and between which were very weak (mean r=0.08). In a principal components analysis, these two sets composed two orthogonal dimensions. One set or dimension of essentialist beliefs corresponds beautifully to the idea of “natural kinds” discussed by philosophers and characteristic of laypeople’s intuitive understanding of living kinds. Categories scoring high on this dimension were judged to be immutable, natural, historically invariant, sharply bounded, and underpinned by necessary or defining features. In short, this factor indexes the degree to which a social category is understood to be specieslike. Gender, racial, and ethnic categories best exemplified natural kinds in the study. Categories scoring high on the second factor were judged to be homogeneous, inductively potent, identity determining, and based on shared inhering properties; that is, to have an underlying “real-ness”. This set of elements closely approximates D.T.Campbell’s (1958) concept of entitativity: homogeneity approximates his coherence, informativeness and identity-definingness or exclusivity approximate meaningfulness, inherence corresponds to the sense of underlying or reified realness that entitative categories possess. Unlike Campbell’s Gestalt account of entitativity, however, this tight ensemble of beliefs suggests that entitativity cannot be understood purely in terms of “perceptual” properties—homogeneity and so forth— but is grounded in ontological beliefs in the inhering, core, inductively potent essencelike properties that underpin them. In any event, we believe that this second dimension of essentialist thinking indexes what Hamilton, Sherman, and Lickel (1998) dub the “entitativity continuum,” although we recognize that it may only do so in regard to social categories, given that entitativity may have different meanings and determinants in different types of group (Sherman, Hamilton & Lewis, 1999). According to our participants, homosexuals, AIDS sufferers, and Jews best exemplified entitative categories. (Jews were also judged to exemplify natural kinds reasonably well, like other “ethnic” groups). Were the abstract belief dimensions associated with category devaluation, as expected, so that shared beliefs about the ontology of social categories have implications for the stigma attached to them? The answer was a clear yes. The entitativity dimension was significantly negatively correlated with status (r=−0. 36), and the more devalued member of each category pair was generally judged to be the more entitative. This negative association was especially strong for categories understood to be natural kinds; in categories falling above the mean on the natural kind factor, the entitativity dimension correlated a massive −0.69 with status, compared to −0.16 for categories judged to be less natural. The two dimensions of essentialism interact, so that when social categories are essentialized in both ways, they are especially likely to be stigmatized. Study 2: Individual Differences in Essentialist Beliefs and Prejudice Our simple first study shows how shared essentialist beliefs about social categories are associated with one another and with stigma. We next sought to determine whether this structure and this association with category devaluation held up at the individual difference level. That is to say, we asked whether people differ in their tendencies to essentialize social categories in these two distinct ways, and whether such differences are associated with prejudice. Gordon Allport (1954), the father of the psychology of prejudice,
54
THE PSYCHOLOGY OF GROUP PERCEPTION
wrote compellingly that they would, almost five decades ago. Prejudiced people tend to hold what he called a “belief in essence” about the inhering properties of category members: There is an inherent “Jewishness” in every Jew…. The “soul of the Oriental,” “Negro blood,” Hitler’s “Aryanism,” “the peculiar genius of America,” “the logical Frenchman,” “the passionate Latin”—all represent a belief in essence. A mysterious mana (for good or ill) resides in a group, all of its members partaking thereof, (p. 169) Accompanying this belief in essence was, Allport said, a cognitive style favoring what he called “monopolistic,” undifferentiated, dichotomizing, and rigid categories. Researchers have found associations between styles such as these (e.g., ambiguity intolerance, rigidity, authoritarianism) and prejudice, but have not examined essentialist beliefs per se. We conducted a simple individual differences study (Haslam, Rothschild, & Ernst, 2002), in which 81 American undergraduates completed eight of the nine essentialist belief items (excluding the item concerning exclusivity or “identity determiningness”) about blacks, women, and gay men. They all also completed self-report measures of antigay attitudes (Attitudes to Gay Men scale [Herek, 1988]), sexism (Modern and Old-Fashioned Sexism scales [Swim, Aiken, Hall & Hunter, 1995], Ambivalent Sexism Inventory [Glick & Fiske, 1997]) and racism (Modern and Old-Fashioned Racism scales [McConahay, 1986], Racial Attitudes Questionnaire [I.Katz & Hass, 1988]). One subgroup completed two popular measures of prejudice-related personality—the Right Wing Authoritarianism scale (Altemeyer, 1988) and the Ambiguity Tolerance scale (MacDonald, 1970)—and another completed a variety of alternative measures of sexism and racism to assess the generality of possible associations between essentialist beliefs and prejudice. The findings were noteworthy in several respects. First, the structure of individual differences in essentialist beliefs, for all three categories of interest, exactly replicated the structure observed in the original study of shared beliefs. Principal components analyses yielded identical, orthogonal natural-kind and entitativity factors for beliefs about blacks, women, and gay men (cf. N.Haslam, 2002, for closely related factors underlying beliefs about depression). Second, just as Allport seems to have been mistaken in implicitly maintaining the unitariness of essentialist thinking, he also seems to have been wrong to present essentialist thinking as a cognitive style that is consistent across social categories. Our participants differed markedly in the degree to which they essentialized the three categories, rather than being across-the-board essentializers. The tendency to understand gay men as a natural kind was correlated only .08 and .12 with tendencies to naturalize blacks and women, and the corresponding cross-category correlations for the entitativity factors were .14 and .23. Essentialist beliefs appear to represent consistently structured modes of thought that can be deployed in quite category-specific ways. A third way in which Allport’s and others’ intuitions about links between essentialist thinking and prejudice seem to need correction is that essentialist beliefs about women and blacks were strikingly unsuccessful in predicting sexism and racism. Across multiple measures of these prejudices, no essentialist belief factors or individual items correlated significantly with any prejudice scale. On the other hand, both prejudices were strongly predicted by authoritarianism and ambiguity intolerance. This picture was reversed for antigay attitudes. Here the essentialist belief items had a very powerful association with scores on the prejudice measure R2=0.49), much of it due to entitativity-related items. In regression analyses the essentialist belief items were associated with antigay attitudes independently of its established personality predictors, together with which they accounted for virtually all of the reliable variance in antigay attitudes R2=. 76). Nevertheless, although these results indicate that individual differences in essentialist beliefs may be
ESSENTIALISM AND ENTITATIVITY
55
associated with category devaluation, as predicted, even here the relation between essentialist beliefs and prejudice was complex, with some essentialist belief elements being positively associated (i.e., inductive potential, discreteness, historical invariance) and others negatively associated with antigay attitudes (i.e., naturalness, immutability). Study 3: Manipulating Essentialist Beliefs The two studies sketched offer some convergent evidence regarding the structure and correlates of essentialist belief, but their evidence is strictly correlational and static. Ideally, to establish the psychological reality and importance of these social ontologies, it would be possible to show that these beliefs are bound together dynamically. To do so, we need to manipulate beliefs and observe the implications for other beliefs. That is, when one essentialist belief is altered, do others change in concert, indicating a set of beliefs that are dynamically interconnected? To test this sort of claim we obviously need to do an experiment and find social categories about which beliefs are modifiable. We decided to investigate people’s essentialist beliefs about mental disorders (N.Haslam & Ernst, 2002) for three primary reasons. First, as a purely methodological matter, if beliefs are to be manipulated they must be manipulable, and folk beliefs about psycho pathologies tend to be rather promisingly ambiguous, fluid, and nonconsensual. Second, and most importantly, essentialist beliefs about mental disorders have yet to be investigated empirically despite some theoretical attention (e.g., Scadding, 1996), but represent foci of professional disagreement among partisans of the many conflicting orientations toward the study of psychopathology. The “medical model” (e.g., Guze, 1992), for example, can be construed as an understanding of mental disorders as natural kinds. According to the model, disorders are disease entities that are biologically based (i.e., naturalness), have specific etiologies (i.e., defining or necessary causal properties), are bounded categories that are discontinuous with normal psychological varation (i.e., discreteness), can be palliated but generally not cured (i.e., immutability), and are fundamentally invariant across cultures and times (i.e., universality, historical invariance) (N.Haslam, 2000, in press). In Fábrega’s (1997) words, biomedicine offers an “ontological theory of disease” that postulates “entities that show…common features regardless of the person’s unique characteristics,” that “have specific causes, manifestations, and courses,” and that “entail… natural objects and natural processes” with a “panhuman relevance” (p. 133). As a third motivation for our focus on mental disorders, essentialist beliefs may have important implications for the stigmatization of the mentally disordered. Consequently, we conducted a study to assess the structure of essentialist beliefs about mental disorders and to determine experimentally whether it corresponded to the previously replicated structure. We developed a questionnaire in which five essentialist beliefs (naturalness, discreteness and immutability from the natural kind set and homogeneity and inductive potential from the entitativity set) were individually manipulated in relation to five varied disorders. The methodology was based on one used previously to investigate entity vs. incremental theories of personality (Chiu, Hong & Dweck, 1997). Fiftyeight participants read vignettes in which a disorder was briefly described, a controversy among experts over one of the beliefs was laid out (e.g., disagreement about whether the disorder had a biological or an environmental basis), and then new, supposedly conclusive scientific evidence in favor of the essentialist position was summarized. Participants rated the plausibility of this evidence, and then the eight essentialist belief items used in the previous study. These items were rated on a seven-point scale in which the midpoint represented the judgment that the evidence would not change their preexisting belief about the disorder, and the scale endpoints indicated change in beliefs in the essentialist or antiessentialist direction. (Additional analyses confirmed the plausibility and power of our manipulations). Each participant read and rated five
56
THE PSYCHOLOGY OF GROUP PERCEPTION
vignettes, one for each disorder and for each manipulated belief, combined in a Latin square design with order of vignettes randomized. If essentialist beliefs cohere dynamically as a set, representing an integrated mode of thinking, participants should infer other essentialist beliefs when one is induced by experimental manipulation. Moreover, to the extent that es sentialist beliefs conform to the two-component structure obtained previously, we would expect a preferential pattern of inferences: Participants should most strongly infer natural-kind beliefs when other natural-kind beliefs are manipulated, and entitativity-related beliefs when other entitativity-related beliefs were manipulated. Both of these predictions were substantially supported. First, participants drew a wide range of essentialist inferences about the disorders when any single belief was manipulated in an essentialist direction. Only 1 of the 35 inferences was contrary to the predicted direction, and 17 were significant at the . 01 level. Second, these inferences were significantly stronger within the natural-kind and entitativity belief sets than between them (p<.001), particularly for the natural kind beliefs. To illustrate, when naturalness was manipulated—by presenting evidence that a disorder had a biological cause—participants strongly and significantly inferred that the disorder was discrete and historically invariant and had necessary or defining properties, but only weakly and nonsignificantly inferred that the disorder was homogeneous, had an underlying or inherent reality, and was inductively potent. This study indicates that essentialist beliefs are dynamically interlinked, and interlinked in a fashion that substantially conforms to the structure obtained in our correlational work. Participants readily drew broad inferences about the nature of mental disorders based on a single, focused piece of evidence and appeared to apply organized social ontologies—natural kind and entitativity beliefs—in doing so. IMPLICATIONS OF THE THREE STUDIES Stated baldly, our three studies indicate that essentialist beliefs about social categories form two distinct and internally coherent sets and have potentially complex implications for category evaluation. The twocomponent structure replicates well, characterizing differences between categories in shared beliefs, differences between individuals in beliefs about three specific categories, and patterns of inference assessed experimentally. It also appears to reflect modes of thinking that are relevant to social categories in multiple domains. There are two primary implications of these findings for theorists and researchers. First, we believe that our findings put the concepts of natural kind and entitativity on a firmer footing in relation to group perception. Until this work was done, we would argue, the two concepts were applied to the study of group perception as useful a priori theoretical notions without any assurance that they mapped onto psychologically real dimensions. As a result, entitativity, for instance, has served in the literature as an explanatory concept, as an operationally defined dependent variable (typically ratings on a simply worded item that emphasizes one aspect or understanding of the concept), or as an experimental manipulation (also based on one selected aspect, normally homogeneity), but not as an empirically derived and described phenomenon. Although we did not initially hypothesize that anything resembling the concepts of natural kind and entitativity would emerge in our work, the dimensions that repeatedly arose corresponded strongly to the ways in which theorists have discussed and formulated each concept. The natural-kind dimension corresponds closely to the elements philosophers and students of folk biology placed under that rubric, and the entitativity dimension’s elements map with little difficulty onto the formulations of Campbell and others. Not only do the findings therefore establish with some confidence that the two concepts refer to psychologically real dimensions of group perception, or real cohering sets of beliefs or judgments, they may
ESSENTIALISM AND ENTITATIVITY
57
also give the concepts more texture, increasing the detail with which we may think about them and highlighting their connections to existing theory outside of social psychology. Second, the findings offer qualified support for existing work on essentialist thinking in the social domain. On the one hand, they support the psychological reality of the two aspects of essentialist thinking, immutability and inductive potential, that were identified as central by Rothbart and Taylor (1992), and also support their use of the concept of natural kind in relation to group perception. On the other hand, they indicate that immutability and inductive potential should not necessarily be seen as privileged facets of essentialism but as belonging to broader sets of beliefs that the studies help to characterize. Indeed, the findings can be taken to imply three cautions to theorists of essentialism and entitativity: (1) they should be careful not to mistake particular essence-related beliefs, such as immutability, for the broader set; (2) they should take care not to treat essentialist thinking as unitary or monolithic, as if all of its elements, such as naturalness and inductive potential, go together; and (3) they should make sure not to use a single broad explanatory concept, such as natural kind or entitativity, as if it were equivalent in meaning to essentialism. RELATIONSHIP OF ESSENTIALISM TO ENTITATIVITY In view of our findings that natural-kind and entitativity-related beliefs compose distinct sets or social ontologies, and in view of the theoretical reflections with which we began this chapter, how should the relationship between essentialism and entitativity be understood? At the outset, we must emphasize that we are restricting our analysis of entitativity to the way the concept applies to social categories and not to other forms of social group, where it may have different meanings and determinants (Sherman et al., 1999). With this in mind, several formulations of the relationship seem possible. First, we might take natural kind beliefs and entitativity to represent distinct subsets of essentialist beliefs, so that entitativity is subsumed within essentialism. Second, we might take natural-kind beliefs to represent the true core of essentialism in the domain of group perception, so that entitativity and essentialism are distinct phenomena with no hierarchical relationship holding between them. Third, we might consider essentialism to represent the conjunction of natural-kind and entitativity beliefs. This third alternative resembles the first in recognizing entitativity as a subordinate component of essentialism, but differs from it in arguing that there is only one way to essentialize a social category. Whereas the first alternative allows for a social category to be essentialized either by being adjudged a natural kind or a reified entity, the third reserves the term for occasions when both ontological assumptions are held. We do not have strong intuitions about which of these three alternatives is the most apt or the most fruitful for further work, and needless to say there is no “true” meaning of essentialism waiting to be discovered. However, the second alternative, which privileges natural kind beliefs as the real basis for essentialism, seems unpromising. There seems little reason to restrict essentialism to natural-kind beliefs— defining features, sharp boundaries, naturalness, immutability, and historical invariance—when many writers have linked the term to entitativity-related elements such as inherence, homogeneity, and inductive potential (e.g., Hirschfeld, 1996; Rothbart & Taylor, 1992). On balance, we favor the third alternative, according to which essentialism represents the conjunction of natural-kind and entitativity elements. This formulation avoids prematurely narrowing the concept of essentialism and also presents it as a socialcognitive phenomenon that is in an important sense unitary rather than divided into two unrelated types. In addition, it avoids the counterintuitive implication that a category taken to be a natural kind but to be without inductive potential or homogeneity—as “white male” is judged to be by many Americans—is understood by them in an essentialist way. Similarly, it avoids treating entitative categories that are not
58
THE PSYCHOLOGY OF GROUP PERCEPTION
taken to be natural kinds as essentialized. This is an important consideration given that many factors besides the assumption of a shared essence can give rise to judgments of entitativity. In short, a conjunctive understanding of essentialist beliefs may be most defensible. Interestingly, with hindsight our earlier work (N.Haslam et al., 2000) offers some support for the basic plausibility of this position. Not only were the social categories often claimed to be prone to essentialist understandings— racial, ethnic, gender, and sexual minorities—judged to be both natural kinds and entitative, but the interaction of these two dimensions strongly predicted evaluative judgments of our sample of categories. Moreover, our mental disorders study (Haslam & Ernst, 2002) indicated that entitativity and natural-kind beliefs are dynamically linked to some degree, in that manipulations of beliefs in each set generated inferences to beliefs in the other set. Our conjunctive understanding of essentialism and entitativity differs in several respects from the relationship of the two concepts proposed by Yzerbyt, Corneille, and Estrada (2001). They view entitativity as an “ecological” or phenotypic aspect of social perception that is connected to an “inferential” essentialist aspect and present a variety of mechanisms by which these aspects are interlinked. This proposal resembles ours in seeing dynamic links between two components of social perception, and in recognizing that essentialism is not simply “perceptual.” However, Yzerbyt et al. see entitativity and essentialism as distinct facets of our unitary entitativity dimension—referring to surface homogeneity and inferred underlying similarity, respectively—and thereby neglect our natural-kind dimension. Although it is elegant and heuristic, we believe that Yzerbyt et al.’s proposal is problematic. The concept of natural kinds (or living kinds) arises in many theoretical discussions of essentialism, and to exclude it seems like a major theoretical omission. Even if the centrality of biological understanding to social essentialism is called into question, our natural-kind factor contains conceptual elements—defining features, sharp category boundaries, immutability, historical invariance—that are fundamental to most treatments of essentialism. Moreover, as noted above, the social categories that are most often considered to be essentialized by the lay public—races, ethnic groups, genders—are clearly prone to some form of naturalized, folk-biological understanding. In short, we believe that to delimit entitativity and essentialism as Yzerbyt et al. propose is to leave out some crucial elements of these concepts as they apply to social understanding. Clearly more theoretical work is needed to resolve disagreements such as these. REMAINING ISSUES There is clearly much work still to be done on essentialism and entitativity, not least on how they should be conceptualized. We are not in a position to present authoritative proposals for how future work should proceed or what the primary research questions should be. However, we do have some concluding thoughts on issues that need further attention. One issue is perhaps mainly terminological, and hence may seem to be of minor importance, although we are not so sure. Although this volume’s title invokes group “perception,” care must be taken not to frame the study of essentialism and entitativity by analogy to sensory perception. Almost by definition an essence is not perceptible, and to have an essentialist understanding of a social category is to make an attribution or judgment about its ontological status and hidden causal underpinnings more than it is to “see” it in a particular way. In this connection, Hirschfeld (1996) has done the field a real service by showing the insufficiency of perceptual factors in accounting for the development of children’s racial understandings. The same goes, we think, for entitativity: to the extent that it involves a reifying attitude to a social category it is not just a perception of how coherent, homogeneous, or consistent the category’s mem bers are, but is also an imposition onto the category of an organized set of beliefs about its nature or basis. To construe
ESSENTIALISM AND ENTITATIVITY
59
essentialism and entitativity as primarily perceptual phenomena—if social perception is taken to be like sense perception—is to omit the ontological component that is fundamental to them, and it may lead researchers to treat familiar notions such as homogeneity as more or less equivalent to them. As a result, researchers may then operationalize these complex concepts in overly simple ways and lose what may be most interesting about them (e.g., using the concept of entitativity to simply recycle the concept of homogeneity). A second issue that awaits further study is how laypeople’s beliefs about the essences of social categories should be understood. Because of the conceptual links between essentialism and folk biology, the common equation of essentialism with some sort of biological reductionism by critical theorists, and the focus of research and theory on social categories that seem to have a biological dimension (e.g., genders, races, sexual orientations), there is a tendency to assume that in the popular mind essences must in some sense be biological or “natural.” This may not be the case, of course. It is not obvious, for instance, that Boyer’s (1993) participants, who treated spiritual leaders as an essentialized “pseudo-natural kind,” understood the special essence of these leaders in even an implicitly biological manner. Although physical features linked to biological variation maybe especially potent triggers of essentialist thinking (e.g., Dasgupta, Banaji, & Abelson, 1999), it is quite possible that some truly essentialist understandings of social categories involve an attribution of some kind of spiritual or otherwise nonmaterial essence. Kofta and Sedek’s (1999) work on “group soul stereotypes” may well be a case of this sort, in which some ethnic stereotypes ascribe essences that are not treated as biological. Some quasi-scientific theories of particular social categories are plainly both essentialist and nonbiological. Lacanian psychoanalysis, for instance, accounts for psychological conditions such as hysteria with abstract psychological structures that are defining, intrinsic, and causally potent; in short, psychological essences. Whether laypeople ascribe nonbiological essences to social categories, and the structure of these non-natural kind understandings, is a matter that merits empirical attention. It is currently an active topic of theoretical debate between those who see essentialist social thinking as involving the application of a “living kinds” module specialized for reasoning about species (Gil-White, 2001), and those who see such reasoning as a domain-general mode of thought rather than being intrinsically biological (Gelman & Hirschfeld, 1999). Resolving this issue is also a daunting task, given people’s normal inability to characterize with any clarity the essence that their apparently essentialist thinking suggests they must believe in. A third issue that we believe needs further attention is the explanation of how essentialist beliefs are associated with category evaluation, specifically stigma and prejudice. It is clear that associations exist, and that attributions will play a role in understanding them (cf. Weiner, Perry & Magnusson, 1988), but it is also clear from our studies, we think, that the associations are complex and inconsistent across categories. Much more theoretical and empirical work will need to be done before we have a full understanding of which essentialist or antiessentialist beliefs are associated with category devaluation and why; which factors determine whether these associations hold across different categories; and to what extent essentialist beliefs promote devaluation rather than devaluation promoting essentialist beliefs. The importance of questions such as these ensures that the study of essentialism and entitativity in relation to social categories will have an interesting future. We hope that researchers and theorists will continue to make contributions to what has emerged as a vigorous field of study, and open themselves to a range of theoretical perspectives in the process.
5 The Mental Representation of Social Categories Category Boundaries, Entitativity, and Stereotype Change MYRON ROTHBART University of Oregon BERNADETTE PARK University of Colorado
If my theory of relativity is proven correct, Germany will claim me as a German and France will declare that I am a citizen of the world. Should my theory prove untrue, France will say that I am a German and Germany will declare that I am a Jew. —Albert Einstein Einstein’s reflections on social classification are instructive for students of intergroup perception. While the most obvious lesson is that his status as an ingroup member depends upon the value of his theory, there are two other ideas implicit in this brief quote that deserve attention, one more obvious than the other. The first simply notes that the same individual may be accurately described by a number of different category labels. Whatever the status of his theories, Einstein can fairly be characterized as a “citizen of the world,” and a “German,” and a “Jew.” Indeed, there is an inclusive or at least partly hierarchical relation among the categories. Citizen of the world subsumes German, and being a Jew overlaps with being German. The assumption that a person belongs to multiple categories is obvious, but the tone of the above quote invokes a second assumption, that membership in one category precludes membership in another. Einstein indicates that Germans will view him either as a German or as a Jew, and the French will view him either as a citizen of the world or as a German. Whereas the realities of the categorical relations involve inclusivity and overlap, the psychological representation of these categories implies exclusivity. It is the psychological representation of categorical relations— and not the external reality of those relations—that constitutes the focus of this paper. The above quote illustrates once again the importance of the stimulus error in psychology, which refers to the assumption that the perception of an object is determined solely by the characteristics of the stimulus— that a stimulus can be perceived in one and only one way. As applied to the study of social categories, the
THE MENTAL REPRESENTATION OF SOCIAL CATEGORIES
61
stimulus error appears to have two corollaries. The first is that a stimulus person’s category membership is necessarily included in the psychological representation of that individual. Put less abstractly, a woman is represented as a woman, a black as a black, a Jew as a Jew, etc. Indeed, there is now good evidence that this assumption is not tenable. Early work by Asch (1940) on the meaning of the category “politician,” and more recent work by Lord and Lepper (1999) on the malleable relation between category and category exemplars has nicely shown that judgments about the nature of politicians, for example, depends strongly on the nature of the exemplars that are momentarily activated in memory, and that the same exemplar can be represented by a number of different category names. It could be argued that the fuzziness of such categories as “politician” or “statesman” allows fluidity in representation, but not so for categories with clearly established boundaries, such as gender or race. If we take as evidence for categorical representation the ability of a category label to activate or “call to mind” the relevant category member, then there is good evidence that women or blacks, for example, who are poor fits to their respective stereotypes are functionally excluded from membership in these categories (Rothbart & John, 1985; Rothbart & Lewis, 1988; Rothbart, Sriram & Davis-Stitt, 1996). Thus, even “obvious” group members may not necessarily be encoded as category members if the overall fit to the stereotype is poor. The important point is that it is not the clarity of the group boundaries (for example, with gender) or the societal importance of the category features (for example, with race) that necessarily determines categorical representation, but the degree of match between the individual and the group stereotype. The stimulus errors second corollary is more subtle. As implied in the opening quote, the psychological representation of category relations may not mirror the logical relations among categories. An individual can belong both to the category German and to the category Jew, but these categories are often represented as exclusive rather than as overlapping. Thus, the second corollary states that there is a tendency for one category to inhibit an individual’s association with another category. Saltz and Medow (1971) found that young children had difficulty in identifying a person already associated with one category (e.g., father) with a second category (e.g., baseball player). Since any hierarchical system implies that an object is associated with more than a single category, and since humans can readily learn hierarchical systems, there are limits to the strength of this tendency toward exclusivity. We nonetheless share Allport’s view that some labels are “exceedingly salient and powerful. They tend to prevent alternative classification, or even crossclassification…‘labels of primary potency’…act like shrieking sirens, deafening us to all finer discriminations that we might otherwise perceive” (Allport, 1954). In our view, categories vary markedly in their ability to “prevent cross-classification,” and we believe it is the “strength” of a category that determines whether it can override or suppress the influence of other categorical information. As we shall argue later in this chapter, the concept of category “strength” is related to issues of the perceived impermeability of category boundaries as well as the informativeness and homogeneity of the category— factors that are, in turn, closely related to issues of group entitativity. More generally, the stimulus error as applied to social categories warns us that the categories associated with a stimulus person are difficult to predict solely from the properties of the stimulus. The perception of individual group members, or of groups-as-a-whole, is far more subjective and malleable than may be apparent. Women, fraternity members, and Jews may or may not be represented as members of these categories, which has important implications for understanding stereotypes and stereotype change. The question addressed by this chapter is: How do we represent, psychologically, the relations among overlapping and inclusive categories, and how does this representation influence our perception of the nature of social categories and the modifiability of stereotypic beliefs associated with those categories? We address this broad question by considering four, interrelated issues: (1) When is a group member included in or excluded from judgments about the group-as-a-whole? (2) How might contextual factors
62
THE PSYCHOLOGY OF GROUP PERCEPTION
affect the inclusion versus exclusion of group members? (3) What does entitativity mean and how does it affect the placement of category boundaries? (4) What is the relation among homogeneity, entitativity, and stereotype change? Ultimately we argue that whether an individual has the power to change a stereotype depends on whether that person is placed inside the group boundaries (i.e., seen as a legitimate member of the group) or is functionally excluded from the group. We argue further that where the group boundaries are placed, or how tightly they are drawn, depends on both the extent to which the group is seen as homogeneous, and the extent to which the group is seen as a meaningful entity. WHEN IS A GROUP MEMBER NOT A GROUP MEMBER? One of the most enduring and fundamental problems addressed by social psychologists is how the mind deals with discrepancy. Theories of cognitive consistency broadly addressed the question of how we reconcile the grouping of social objects of differing affective sign (e.g., Abelson, 1959; Festinger, 1957/ 1962). In the domain of stereotypic beliefs, the question became one of resolving the discrepancy in perception between the attributes of a group and the attributes of a counterstereotypic group member. One solution, proposed by Crocker and colleagues (Weber & Crocker, 1983; Crocker, Fiske, & Taylor, 1984), was that a member who strongly disconfirms the stereotype of the group as a whole is “subtyped”—that is, is (implicitly) given a label that functionally removes them from the group. Although there are ambiguities surrounding the concept of a subtype (which will be discussed later), the basic idea is that the member is implicitly or explicitly given a subordinate label (e.g., “hard-driving female CEO,” instead of “woman”) defining the “sub-type” which functionally isolates the individual from the group. The original evidence for this phenomenon from Weber and Crocker (1983) was based on the finding that group members with many stereotype-disconfirming attributes changed the stereotype less than did individuals with fewer such attributes. Subsequent research by Hewstone and colleagues (Hewstone, 1994; Hewstone, Johnston, & Aird, 1992; Hewstone & Lord, 1998; Hewstone, Macrae, Griffiths, Milne, & Brown, 1994), and by Lord and colleagues (Desforges, Lord, & Pugh, 1997; Lord, Desforges, Fein, Pugh, & Lepper, 1994; Ramsey, Lord, Wallace, & Pugh, 1994; Scarberry, Ratcliff, Lord, Lanicek, & Deforges, 1997) has replicated and extended these findings (see Richards & Hewstone, 2001, for a review of this work). The general idea behind subtyping is that the recategorization of a poor-fitting group member into a subordinate cluster has the effect of reducing the impact of that group member on the overall group impression. A different line of research, focusing on the factors that lead to changes in perceived intragroup variability, was conducted by Park, Ryan, and Judd (1992). By having subjects focus on the attributes of the subgroups that make up the group as a whole, they were able to show in both experimental and correlational studies that attention to subgroups increased the degree of perceived intragroup variability and decreased the strength of the stereotype. Note that at some level this is exactly the opposite of what would be predicted by the subtyping principle. The latter suggests that focusing on a subordinate category should lead to exclusion of the poor-fitting members, resulting in greater perceived intragroup homogeneity and increases in stereotypic perception. Clearly, the reasoning behind the subtyping principle and the findings of Park et al. on subgrouping appear to be in clear opposition and a resolution of this discrepancy was needed. Maurer, Park, and Rothbart (1995) argued that subtyping and subgrouping were two functionally different processes, and although both involve the establishment of subordinate categories, they lead to different effects on the perceived strength of the stereotype. Subjects were given information about a group of 16 stimulus persons involved in a Big Brother program, where the group members varied in typicality. One group was given this information without any instructions (the control condition) and then asked to make typicality judgments about individual group members as well as judgments about the group as a
THE MENTAL REPRESENTATION OF SOCIAL CATEGORIES
63
whole (including judgments of stereotype strength and intragroup variability). A second group was given subtyping instructions to sort the group members into those who fit and those who do not fit the image of the group before making the same judgments as did the control condition. A third group was given subgrouping instructions to sort the members into as many piles as they wished, trying to minimize differences within subgroups while maximizing differences between subgroups before making their judgments. Importantly, the stimulus materials were constructed so that, in addition to a disconfirming subgroup, the confirming group members could be organized into four plausible subgroups, each of which confirmed the group stereotype in a distinct and idiosyncratic manner. Compared to the control condition, the subtyping condition had a more stereotyped and homogenous image of the target group, while the subgrouping condition had a less stereotyped and more heterogeneous image. Typicality ratings indicated that identical disconfirming group members were viewed as more atypical in the subtyping condition but less atypical (i.e., more typical) in the subgrouping condition. Based on this analysis, it was concluded that the atypical group members were functionally excluded from the group representation in the subtyping condition but were included in the subgrouping condition. There are two potentially important implications of this research. One is that the perception of typicality is influenced not only by the “computation” of matching features between category and exemplar, but also by the context of judgment. Context can determine whether subjects do or do not include the atypical members in their implicit calculations of group impressions, and this is an important phenomenon that needs to be better understood. In this research the subtyping and subgrouping instructions had very different effects on how subjects thought about the relation of group members to the group as a whole. The relatively simple dichotomous judgments (good fit vs. poor fit) required by the subtyping instructions may have led subjects to view group members in a unidimensional way, promoting an exaggerated difference between the attributes of good- and poor-fitting members (in comparison to the control condition). In contrast, the more complex subgrouping instructions, which encouraged subjects to examine similarities and differences between and among individual stimulus persons and between and among subgroups, may have had at least two important consequences. First, subjects may have been led to realize that each stimulus person was a complex set of attributes, some of which fit and some of which did not fit the stereotype (unlike the global “good fit vs. poor fit” in the subtyping condition). In contrast to the subtyping instructions, the multiattribute nature of the stimulus persons was emphasized, leading to more moderate judgments of atypicality (cf. Judd & Lusk, 1984). Second, whereas the judgments under subtyping instructions may have led subjects to think of atypical members as outside the group, or as nonmembers of the group, the subgrouping instructions may have encouraged subjects to make discriminations within the context of an activated superordinate category. To state this somewhat differently, there is a difference between sorting a list of occupations into those that fit or do not fit, for example, our stereotype of women (the subtyping instructions) versus a task in which we classify a list of occupations, all engaged in by women, into subcategories (the subgrouping instructions). In the former case, the atypical exemplars may be dissociated from the superordinate category, and in the latter case may be actively associated with that category. The second general implication of this work is the importance of implicit group boundaries. One way to think about the Maurer et al. (1995) results is that subjects in the subtyping condition are implicitly drawing the group’s boundaries in a way that is different from subjects in the subgrouping condition. For the former the boundaries appear to exclude atypical group members, whereas for the latter the boundaries appear to be drawn in a more inclusive manner. Although the concept of “implicit group boundaries” remains vague, it has precedence in G.Allport’s (1954) concept of “refencing” (referring to the “redrawing” of group boundaries to place poor-fitting group members outside the category) and in Lewin’s (1936, 1948/1997) treatment of psychological barriers. In both cases, it is argued that one function of psychological boundaries
64
THE PSYCHOLOGY OF GROUP PERCEPTION
is to seal off or isolate one region of the “life-space” from another. The isolated regions may be individual group members isolated from the group as a whole, one social category isolated from another, the ingroup isolated from the outgroup, or any related concepts or ideas that remain isolated or inaccessible from one another (cf. Nissen, Ross, Willingham, Mackenzie, & Schacter, 1988). In the Maurer et al. (1995) paradigm, the subtyping and subgrouping participants exhibited their respective patterns of organization presumably because of the cognitive perspective they were instructed to adopt. In principle, both subtyping and subgrouping can result from a particular cognitive orientation as information is processed (i.e., fit-no-fit vs. multiattribute grouping). A priori, motivational factors seem to be more likely to operate in subtyping than in subgrouping. The subtyping model was developed to explain why individuals maintain stereotypes in the face of disconfirming information (Rothbart, 1981). To the extent that one wishes to hold on to a set of beliefs about a group, subtyping is a means for doing so. In contrast, subgrouping seems less likely to be driven by motivational concerns. What Is Subtyping? Before considering some of the factors that influence boundary placement, we would like to examine more precisely the issue of subtyping, because it may clarify at least two different ways of thinking about the representation of subor dinate and superordinate categories. The original work by Weber and Crocker (1983) and Crocker, Fiske, and Taylor (1984) suggests that subtyping is the process whereby a poor-fitting member of a group is associated with a label that relegates the category member to a subordinate class within the larger group. That is, the hard-driving female CEO is a subcategory nested within the superordinate category of women. According to this view of subtyping, poor-fitting members are viewed as a coherent subcategory and are given a subordinate label. A somewhat contrasting approach was taken by Rothbart and John (1985). They suggested that a poor-fitting member could be “functionally isolated” from the superordinate category, with or without the application of a label, but they were less than clear as to whether the applied label was a logical subtype or an alternative category. The first issue, then, is simply that a poor-fitting member of a group can be isolated or compartmentalized from the group without necessarily applying any label to that member. The second issue, and the one of more general interest for this chapter, is the nature of the applied category. Rothbart and John (1985) used the example of a black doctor (putatively a poor fit to the category blacks) and suggested that this person might be stored with the label doctor, rather than the label black. Although they could have been clearer in their argument, we believe that in contrast to the examples above from Crocker et al. (1984), in which “executive” was really considered to be “executive woman,” in the case of Rothbart and John, they view doctor as an alternative to the category blacks, not as a subclassification within blacks. More generally, when a label is applied to a poor-fitting category member, the label can in principle be of two sorts: (1) a subcategory label which is represented as nested within the superordinate category (i.e., a nested representation), and (2) a category label in which the classification is represented as independent from or as antithetical to the superordinate category (i.e., an independent representation). This distinction may be meaningful since models of category activation would suggest that activation of the subtype in the first case (nested representation) should activate—at least to some degree— the superordinate category, whereas in the second case (independent representation) one would expect either no activation and possibly active inhibition of the superordinate category. Exemplar models would then predict more change in the superordinate category for the nested than the independent representation. More generally, however, the second view, more consistent with the proposals of Rothbart and John, again deals with the difference between “objective” category structure (where, for example, blacks and doctor are
THE MENTAL REPRESENTATION OF SOCIAL CATEGORIES
65
overlapping sets) and represented category structure (where blacks and doctor may be mutually inhibiting). In this instance, the highly discrepant features associated with the two categories introduce a functional boundary between those categories. We now turn to two other factors which influence boundary placement. CONTEXTUAL FACTORS THAT AFFECT THE INCLUSION VERSUS EXCLUSION OF GROUP MEMBERS Language Language may play a particularly important role in defining group boundaries. While touring a Northern European city, one of us came across a plaque memorializing the city’s Jews killed during the Nazi period. The wording was simple and direct, saying, “The Jews of [this city] were rounded up by Nazi forces on [specific date] and sent to the extermination centers in the East.” Later on that same trip, on the banks of the Tiber in Rome, another plaque conveyed the same basic information, but referred to the victims as “our Jewish brothers and sisters.” Although referring to Holocaust victims as “the city’s Jews” or as “our Jewish brothers and sisters” may seem like a negligible difference, the difference in emotional impact was not. The former refers to the victims as “they,” placing a boundary between the observer and victim, while the latter refers to the victims as “we,” and by including them in the same category places the boundary around both the observer and victim. The emotional consequences of the pronominal usage of “we” and “they” have been nicely demonstrated by Perdue, Dovidio, Gurtman, and Tyler (1990). The classic work by Cialdini, Borden, Thorne, Walker, Freeman, and Sloan (1976) found subjects more likely to describe a home team’s victory as one achieved by “we” and “us,” while a defeat was suffered by “they” and “them.” In this case language usage can be considered a marker for identification, which is another way of saying that the implicit boundary encloses the self and a favorable group, but separates the self from an unfavorable group. Is language merely the expression of the underlying psychological representation, or does language itself influence that representation? The answer, in our view, is that both are possible. In the Cialdini study, we suspect it is the implicit psychological boundary that determines the choice of language, but it is entirely possible, as in the case of the European traveler cited above, for the choice of language to determine the implicit boundaries. Certainly the language of dehumanization, used most often in wartime, is one that places the enemy in a category outside the human family (e.g., vermin, reptile, parasite, virus, etc., Keen, 1986). Indeed, it is precisely the connotative imagery associated with such descriptions that magnifies the “psychological distance” between ingroup and outgroup. Intergroup Context Sherif, Harvey, White, Hood, and Sherif (1988), in their classic work on the Robber’s Cave, proposed that the “functional relations” between groups determine the nature of intergroup experience. Put most simply, a competitive relation fosters hostility while cooperation promotes harmony. Rabbie and Horwitz (1969) lay the foundation for reinterpreting Sherif’s findings, suggesting that shared fate, rather than competition or cooperation as such, may be the key ingredient of intergroup attitudes. More specifically, as groups move from competition to cooperation, the group itself is redefined; that is, the group boundaries are redrawn based on shared fate. In a competitive relation, shared fate exists only within each of the opposing groups, and competition ensures different (indeed, opposing) fates for the two groups. Under cooperative conditions, however, shared fate becomes common to the members of both (previously established) groups, and in the
66
THE PSYCHOLOGY OF GROUP PERCEPTION
Robbers Cave Study it may be the changes in shared fate that functionally redefine the group boundaries to create one group where previously there had been two (cf. S.L.Gaertner, Dovidio, Nier, Ward, & Banker, 1999; S.L.Gaertner, Mann, Murrell, & Dovidio, 1989; L.Gaertner & Schopler, 1998). Ingroup versus Outgroup, or Ingroup and Outgroup? The psychological importance of the distinction between ingroup and outgroup embodies two seemingly opposite qualities. On the one hand, while the exact positioning of a boundary between ingroup and outgroup along a given dimension can be arbitrary, once the placement has been made the boundary marker takes on an absolute, almost “natural kind” quality. On the other hand, although on a given dimension the specific boundary takes on undue significance, a change of dimension on which two individuals (or groups) are compared allows them to be in the same (ingroup) category or in opposing (outgroup) categories. This principle is implicit in the opening quote from Einstein, but is most apparent in Allport’s description of a character from an H.G.Wells novel (G.Allport, 1954). The narrator is a systematic botanist who despises plant physiologists (outgroup) unless the latter is viewed as a botanist (ingroup). Similarly, the narrator despises physicists (another outgroup) unless the latter is represented as a scientist (another ingroup). Similarly he despises psychologists unless the latter is reclassified as “educated” (still another ingroup), and despises workingmen (still another outgroup) unless recategorized as Englishmen (yet another ingroup). Given the enormous flexibility in the choice of dimensions on which two individuals or groups can be compared, it seems possible to view any two individuals or groups as sharing a common category or as on opposite sides of an intergroup boundary—depending entirely on the level of classification chosen. At the lowest level, it is possible to differentiate between groups that are on most attributes highly similar, such as plant physiologists versus plant taxonomists (cf. Freud’s “narcissism of small differences”), and at the highest (and more abstract) levels, to see commonality among otherwise diverse categories (human beings, inhabitants of planet Earth, living things). The actual determinants of the preferred level of classification is a difficult issue (and well beyond the scope of this chapter), but is generally thought to be dependent upon the context of intergroup interaction and the needs of the individual (Brewer, 1993). The implicit or explicit presence of the outgroup, for example, increases the salience of ingroup attributes, often blurring otherwise important distinctions within the ingroup (Doise, 1969). And as Sherif et al. (1988) has shown, redefining the functional relations between groups through interdependence can lead to a breakdown in the ingroupoutgroup distinction and what Gaertner and Dovidio (S.L.Gaertner, Dovidio, Nier, et al., 1999; Gaertner et al., 1989; see also Turner, 1985, 1987) have called the “one group from two” phenomenon. Thus far we have considered contextual variables—variables external to the fit between category and category member—that influence whether a group member is categorized within the boundaries of a group. These contextual variables include the mental organization of group members, the use of language, the intergroup context in which perception occurs, and the way in which the latter two factors can influence the “level” of categorization that is used in perceiving individuals and groups. We turn now to the question of how categories differ among themselves, how these differences affect boundary placement, and how this in turn relates to questions of stereotype change. ENTITATIVITY AND ITS ROLE IN DETERMINING GROUP BOUNDARIES It is commonly assumed that contact with individual group members who disconfirm the group stereotype is a powerful method for modifying our beliefs about groups (Pettigrew, 1998). However, a major impediment to such contact-induced change is the failure to represent poor-fitting individuals as members of the group.
THE MENTAL REPRESENTATION OF SOCIAL CATEGORIES
67
The question raised earlier in this chapter is whether these poor-fitting group members are included within the superordinate group in a way that decreases stereotypic perception (subgrouping) or are functionally excluded from the group in a way that allows the stereotype to continue unperturbed (subtyping, either via a nested or independent representation). Although one determinant of category inclusion is the overlap of features between category and category member, we have already considered evidence showing that other factors influence this process (Kunda & Oleson, 1995, 1997; Maurer et al., 1995). Whether or not disconfirming exemplars will be included within a category depends at least in part on the nature of the category boundaries. If the group is viewed as homogeneous, with narrow boundaries, and has clear criteria for group membership, it may be much more difficult for a poor-fitting member to be represented within the group. If the group is more heterogeneous, with broader boundaries, and with fuzzier criteria for membership, information that diverges from the group norm is more likely to be incorporated into the group impression. The nature of group boundaries is strongly related, in our view, to the issue of group entitativity. When Campbell (1958) coined the term entitativity to describe the degree of perceived coherence of social groups, he listed four principles of perceptual grouping basic to the visual domain and attempted to indicate how these principles could be quantified in the physical and social worlds. The four Gestalt grouping principles were proximity, similarity, common fate, and good continuation. In our view, the principle of proximity in determining the perceived entitativity of social groups is of marginal importance. Although proximity can serve as a significant marker for entitativity (e.g., Californians), clearly it is possible to perceive a group that is physically quite dispersed as highly entitatative (e.g., homosexuals). The principles of similarity, common fate, and good continuation are, in our view, the most central to understanding the perceived entitativity of social groups, and these will be considered in some detail. The principle of similarity, which we interpret as grouping by common attributes, we consider basic and necessary for the perception of entitativity. These shared attributes might be physical (Dasgupta, Banaji, & Abelson, 1999), emotional, cultural, or societal and may reflect uniform adaptation to common economic conditions. In social psychology the lingua franca for summarizing behavioral characteristics is that of personality trait descriptors. Thus, if many of the members of a group are seen as aggressive, driven, and pushy, this will lead to a heightened sense of perceived entitativity. In contrast, if the group is seen as possessing these characteristics on average, but with large individual differences across group members in the manifestation of these traits, entitativity will be lower. The principle of similarity can also be conceptualized as simply the extent to which the members of a group are seen in a global sense as similar to one another, with high intragroup similarity indicating high entitativity. We believe that the principle of similarity is a relatively weak determinant of entitativity in comparison to the other grouping principles, in particular that of shared purpose. The second component of entitativity concerns common goals or intentions among the members of a group (Abelson, 1998; O’Laughlin & Malle, 2002), which corresponds closely to Campbell’s principle of grouping by common fate. This is a powerful grouping principle in visual perception, especially for the perception of movement, and seems particularly appropriate to the perception of “movement” or “direction” of social groups. Groups are frequently organized around the achievement of goals, and have a purpose or intention, whether stated or implicit, for their existence. We might say that the group has a strong sense of agency. The truer this is, the more a group can identify clear, definable goals as a desired end state, the more entitative the group will be viewed. Sometimes groups exist for the purpose of achieving some specific, short-range goal, such as the “National Committee for the Election of Al Gore to the Presidency.” Other times the group works toward achieving a general set of long-range goals, but intentions are nonetheless clearly present. Such is the case for groups like the Society for the Prevention of Cruelty towards Animals, the Democratic Party, and the National Organization of Women.
68
THE PSYCHOLOGY OF GROUP PERCEPTION
The third principle of grouping, good continuation, Campbell considers relevant to the definition of group boundaries. Although not well developed in Campbell’s 1958 article, we consider the issue of boundaries to be important. Thus far we have considered two principles, similarity and common fate, which can be thought of as the “glue” that binds the members of a group to one another—either by common attributes or common goals. Group boundaries, in contrast to the dimensions of within-group commonalities, define the separation between the group and the surrounding region. Group boundaries, either sharp or fuzzy, determine to a large degree who is, or is not, considered a category member. Gender is a good example of a classification based on reasonably clear criteria, whereas the apparently clear distinction, say, between the nosological categories of depression and anxiety now appear to be quite fuzzy (Watson & Tellegen, 1985). Group boundaries also seem to differ on the dimension of permeability—the ease or difficulty of entering or leaving the category. Some categories, such as “cashier,” “progressive,” “student,” or “American” are easy to enter, while others, such as “neurosurgeon,” “hero,” “Nobel Laureate,” or “Japanese” are quite difficult. Although there may be an asymmetry between entering and leaving a group, we suspect the two are correlated. In our view, groups that are hard to enter are also hard to leave, and that means that even though an individual may no longer be a member of a group—by some logical or legal criterion—they may nonetheless be considered to retain the important attributes of the group. Someone who retires as a neurosurgeon or repudiates his or her Nobel Prize is likely to be perceived as retaining the characteristics of their former group. This component of good continuation, that is, impermeability of group boundaries, is related to the notion of essentialistic perception. Entitativity and Essentialistic Perception It is difficult to examine questions of homogeneity, agency, group boundaries, and entitativity without considering, however cursorily, the relation between entitativity and the perception of essence. We have already considered the relations among common attributes, common goals, and entitativity. The ascription of essential properties to a group is, in our view, complex, and cannot be predicted solely from a group’s perceived homogeneity and the perceived impermeability of its boundaries. We define essentialistic perception as the ascription of fundamental, basic, or essential dispositions to a group that are regarded as highly enduring and transmitted across time and space through the “mysterious mana” of biology or culture. Rothbart and Taylor (1992) noted that “natural kind” categories (e.g., gold, raccoons, roses) are assumed to have essential or defining features, and that some social categories are responded to as if they were natural kinds. They argued further that only some groups would be perceived as natural kinds— and possessing essential features—and those would be categories with a high degree of inductive potential and perceived unalterability. Inductive potential refers to the ability to make strong inductive inferences from one member to all other members, and it is manifest in a high degree of perceived intragroup homogeneity. Second, the essential group characteristics are perceived as unalterable. While the appearance of a group member can be altered, the underlying essence is assumed to remain constant. These two characteristics are viewed as basic to our perception of natural-kind categories. Rothbart and Taylor speculated that “whereas social categories are in reality more like human artifacts than natural kinds, they are often perceived as more like natural kinds than human artifacts” (1992, p. 12). Category labels with a biological flavor, such as race, ethnicity, breed, caste, and so on—whatever their ontological status—may take on a natural kind quality, whereas mundane occupational categories such as cashier, waiter, or cab driver clearly would not be expected to do so. Once again, language may prove to play an important role in essentialistic perception. In a very interesting recent study, Gelman and Heyman (1999) examined the inferences made by five- to seven-year-
THE MENTAL REPRESENTATION OF SOCIAL CATEGORIES
69
FIGURE 5.1. Hypothesized relations among shared goals, shared attributes, entitativity, and essence
old children when presented with information about a hypothetical person. In half the cases a stimulus person was described by a verbal predicate (“Rose eats carrots whenever she can”) and in half the cases, by a noun label (“Rose is a carrot-eater”). Subjects made far stronger inferences about Rose, over time and across settings, when she was described with a noun label than with a strong verbal predicate. Gelman and Heyman speculated that the use of such noun/category labels promoted the use of natural-kind thinking on the part of subjects. In our view, the ascription of essential properties to a group is highly determined by social context, including the network of existing social beliefs and patterns of intergroup conflict (cf. the “social constructionist” view described by Gil-White, 2001). We view essence as nested within entitativity, as depicted in Figure 5.1. This figure makes clear several of our working hypotheses, and is offered in the spirit of helping to clarify the often confusing relations among some difficult concepts—with the full expectation that it will be altered and clarified as research and theory progresses. Figure 5.1 makes a number of points. First, note that shared attributes and shared goals are partially, but not wholly, overlapping. They are independent constructs, although related. Both enhance the perception of entitativity, but neither is sufficient to produce it—that is, there are many groups high on shared attributes or shared goals that are not seen as particularly entitative. Second, while most groups that are high on entitativity are also high on shared goals, not all are. In contrast, for shared attributes, we see a necessary relation such that any group seen as high on entitativity is also seen as high on shared attributes. Hamilton (D.L.Hamilton, personal communication, 1996) takes issue with this last claim, arguing that there are in fact some groups high in entitativity that are not seen as high on similarity (typically such groups are seen as high on shared goals). Hamilton would not have the entitativity circle fully contained within the shared attributes circle. This disagreement may in part be one of semantics, and it may be due to our perception that high entitativity can cause a perception of shared attributes (see below).
70
THE PSYCHOLOGY OF GROUP PERCEPTION
Third, Figure 5.1 indicates that groups perceived as having essential features are nested within, and are a subset of, groups perceived as highly entitative. That is, groups that are high in essence are also highly entitative, but there are many highly entitative groups that are not imbued with essential features (e.g., taskoriented groups). This latter observation brings us to the general issue of the direction of causality. Is it the case that shared attributes, shared goals, and essence all determine or lead to high entitativity, or are there conditions under which a group is seen as highly entitative, perhaps due to the sorts of contextual factors we described earlier in the chapter, and that this in turn causes the group to be seen as sharing attributes and goals? Our belief is that both can and do happen. While most of our discussion and in fact most of the work in the existing literature is framed in terms of the factors that predict (i.e., cause, determine) entitativity, an interesting psychological phenomenon occurs when groups that are seen as entitative, perhaps due to shared attributes, come to be viewed as high on the other correlates of entitativity, for example, shared goals. Perhaps the most extreme example of this phenomenon was the Nazis’ ascribing to Jews the goal of world domination. Finding it difficult to reconcile the prominence of Jews in both capitalist and anticapitalist endeavors in the late 19th and early 20th centuries, Nazi ideology concluded that the intention common to all Jewish activities was world domination, whether through capital ism or Bolshevism. A contemporary example is provided by conservative religious groups who frequently refer to a “homosexual agenda,” as if there were a unified, strategic plan for recruiting or seducing young children into homosexuality. It is possible that references to a Jewish plan for domination or to a homosexual agenda are simply figures of speech for summarizing what is really being perceived as a disparate effort by individuals to achieve, without organization, a value held in common (cf. O’Laughlin & Malle’s, 2002, distinction between group members’ prevalent reason vs. a group’s joint reason). Although this interpretation should be taken seriously, the choice of language, as discussed earlier, is important, and summarizing a group’s attributes with the language of intentionality is likely to have effects that strongly enhance perceived entitativity. It is also worth noting that the attribution of a common intention may be particularly difficult to disconfirm given its inferential character. What overt behavior would validate the absence of intent (cf. Reeder & Brewer, 1979; Rothbart & Park, 1986)? Given the difficulty of disconfirming a latent intent, it is not surprising that such attributions, when applied to groups, can give rise to conspiratorial explanations for important events. Is Entitativity Anything More than Homogeneity? There is now a large body of knowledge regarding perceived homogeneity (Mullen & Hu, 1989; Ostrom & Sedikides, 1992; Park, Judd, & Ryan, 1991; Park & Rothbart, 1982) and it is useful to examine the relation between entitativity and homogeneity. In the literature, homogeneity has been defined with respect to specific attributes or overall similarity of the group. That is, groups perceived to be high in homogeneity are groups where a large proportion of the members share common attributes (e.g., most actors are extroverts), or where the members are seen as very similar to one another. In our analysis this is the first component of entitativity, shared attributes. Homogeneity and high entitativity have this in common: they both suggest a high level of shared attributes among group members. The characterization of most members of the NBA as tall, athletic, and competitive and members of the Survivalists as white, politically conservative, and militant contribute to the perception that these groups are both highly homogeneous, and that they are high in entitativity—that is, that they are strong groups. In our view homogeneity and entitativity do not share the other two characteristics ascribed to entitativity —agency or essence. In terms of the literature it is clear that it is similarity in the semantic content of the group stereotype (i.e., similarity of attributes) that is being assessed. The measures developed to assess
THE MENTAL REPRESENTATION OF SOCIAL CATEGORIES
71
perceived homogeneity (Park & Judd, 1990) primarily concern the perceived prevalence among, and distribution of, group members on an attribute dimension, or judgments of the similarity of group members either in general or with respect to these attribute dimensions. Researchers have not conceived of homogeneity with respect to the intentions or goals of the group, and this is one of the characteristics that discriminates between the constructs of entitativity and homogeneity. Consider people who are against gun control, such as many citizens of the Intermountain West. The members of this “group” vary widely in their reasons for opposition to gun control (ideology, personal safety, individual rights, history, and tradition), and yet they likely share a good number of attributes. They tend to be politically conservative, pro-guns, individualistic, and male. In this sense, the group is homogeneous, because there are clear, identifiable attributes shared by a large proportion of group members. However, this is not a particularly entitative group. Beyond the one shared goal of preventing gun control, the members of the group have few other agendas in common. In contrast, consider members of the militant Survivalists group. This group is also high in homogeneity, again politically conservative, pro-guns, individualistic, and male dominated. However, this group is higher in entitativity. The group has numerous shared goals and intentions, in contrast to having only a single goal. In addition to opposing gun control, the group opposes government control of almost any sort, they aspire to an all-white society, and they promote violence as a means for achieving these goals. Any group with such encompassing common goals and intentions will also have attribute dimensions that are shared. But it is possible to have common attributes without having shared goals. Thus, shared attributes (or homogeneity) is not sufficient to produce high entitativity. But adding agency to shared attributes increases the level of entitativity, and adding a component of essence increases it further still. In one of the first papers to reintroduce the concept of entitativity, Hamilton and Sherman (1996, p. 349) made essentially the same argument. In considering the relation between entitativity and homogeneity, Hamilton and Sherman suggested that the two constructs are related in that both are influenced by the perceived similarity among group members. However, they go on to argue that entitativity also includes a sense of shared purpose or goals, shared history, or the interdependence of group members (p. 350), and thus while to some extent overlapping, the two constructs do make independent contributions to the intergroup puzzle. As noted earlier, a point of disagreement concerns whether homogeneity is simply correlated with, or necessary for, perceived entitativity. Other recent literature is also consistent both with our analysis of the defining characteristics of entitativity, as well as the argument that entitativity is more than just perceived homogeneity. In a series of three studies Lickel, Hamilton, Wieczorkowska, Lewis, Sherman, and Uhles (2000) asked participants to rate 40 groups on nine properties that they believed might be related to the degree to which a group is perceived as having groupness. Six of the nine dimensions were all moderately intercorrelated. These included a direct rating of perceived entitativity (to what extent does the group qualify as a group), importance of the group to group members, degree of interaction among group members, common goals, shared outcomes, and perceived similarity among group members. The latter rating is a judgment of perceived homogeneity, and the remaining dimensions (excluding entitativity) are all part of or directly related to shared goals and intentions. Three remaining dimensions were essentially uncorrelated with these six. These included duration of the group, permeability of group boundaries, and group size. Yzerbyt, Rocher, and Schadron (1997) and N.Haslam, Rothschild, and Ernst (2000) have also examined a number of dimensions believed to be the defining characteristics of category “essentialism.” For example, Haslam et al. asked participants to rate 40 social categories on nine dimensions that empirically hung together in two clusters. Haslam et al. argued that the first cluster of dimensions reflected properties of natural kinds and, in particular, reflected the alterability of a category. The dimensions on this cluster
72
THE PSYCHOLOGY OF GROUP PERCEPTION
included perceived discreteness (i.e., the sharpness of the category boundaries), immutability, stability, and necessity of features. The dimensions on the second cluster included informativeness, inherence (whether the underlying reality is biological in nature), exclusivity, and uniformity (inductive potential). Haslam et al. argued that this second cluster is basically entitativity, although, as we have defined entitativity above, aspects of the first cluster tap entitativity as well. Our preferred interpretation of Haslam et al.’s results is that the two dimensions correspond rather well to Rothbart and Taylor’s two dimensions of unalterability and inductive potential, both of which we consider components of essence and therefore both contributors to high entitativity. Finally, in our own research we have examined four dimensions along which groups vary that may contribute to the “strength” of the group (Rothbart, Peters, & Buckingham, 2003). Participants were asked to rate a total of 149 groups on the four dimensions of strength, and on an additional dimension of favorability.1 The four attributes related to strength included a judgment of (1) boundary permeability: the ease of moving in to or out of the group; (2) stability, the likelihood of continuing to possess group characteristics after leaving the group; (3) informativeness, the extent to which category membership provides important information about a category member; and (4) similarity, degree of perceived similarity among group members: A fifth variable, favorability toward the group, was also assessed. Boundary permeability and stability were highly correlated with one another (r(148)=−.71) such that the more difficult it is to move in and out of a group (i.e., the less permeable the group boundaries), the more likely it is that group characteristics would be retained even after moving out of the group. Boundary permeability was correlated moderately with informativeness (r=–.40) and similarity (r=−.36) such that the less permeable the group boundaries the greater the informativeness and intragroup similarity of the category. Stability was correlated with informativeness (r=.63) and similarity (r=.57), and informativeness and similarity themselves were highly correlated (r=.77). All four measures loaded strongly on a single factor that we called category strength. Scores on this factor correlate moderately with favorability (r=−.30) such that stronger categories tend to be less favorable—consistent with the specu lation that essentialistic attributions are more prevalent among unfavorable groups. The ten weakest categories, according to these ratings, were cashiers, gas station attendants, waiters, grocery workers, clerks, janitors, tourists, motorists, telemarketers, and factory workers. Note that each of these categories is extremely heterogeneous; category members share few common attributes (except the category label), there is little or no interdependence among group members, and group goals are minimal or weak. Moreover, for these groups natural-kind status seems completely implausible. The ten strongest categories were members of the KKK, ministers, rapists, Nazis, basketball stars, murderers, terrorists, AIDS patients, Mormons, and doctors. Within each category, the members are very homogeneous, agency and intent is apparent for at least some of these groups, natural-kind attributions are plausible for some, and the groups are predominantly (but not exclusively) very undesirable. Looking within the various dimensions, it is possible to identify groups that are high on both the strength factor as well as similarity, as contrasted with those that are high on similarity but low on strength. This gives us a way to examine our earlier premise that homogeneity is a necessary but not sufficient condition for high entitativity. The five groups scoring highest on both the strength factor and similarity were members of the KKK, terrorists, Nazis, ministers, and basketball stars. The five groups having similarity scores nearly as high but having much more moderate scores on the strength factor were bulimics, airline pilots, sorority members, fraternity members, and Jehovah’s Witnesses. It is striking that the first five groups all have a strong singularity of purpose and exist for the purpose of achieving a common goal. In contrast, within the second set of groups, members share some defining characteristic that makes them similar to one another, but there is not nearly the same sense of singularity of purpose among group
THE MENTAL REPRESENTATION OF SOCIAL CATEGORIES
73
members. Airline pilots, like bulimics and members of a sorority, have no uniform goal they strive to achieve. Note that this is not a perfect analysis for several reasons. Similarity is included as a measure in the strength factor, so obviously it is not independent of scores on the factor. Perhaps more importantly, none of the measures taps the dimension of shared goals or intentions, something we have argued is important in determining the degree of entitativity. Nevertheless, as a very rough analysis, the results are interesting. Even though we did not measure shared goals, this dimension does appear to come through in separating high similarity groups that also differ in terms of strength. IMPLICATIONS OF GROUP HOMOGENEITY AND ENTITATIVITY FOR STEREOTYPE CHANGE Earlier in this chapter we argued how individuals who disconfirm to a stereotype are represented, and how that will determine whether or not stereotype change will occur. Specifically, we argued that when disconfirming instances are subtyped, little or no stereotype change should be expected. Moreover, whether disconfirming instances are subtyped or subgrouped depends in turn on where the group boundaries are drawn, that is, on whether the instance is seen as a group member or not. We are now ready to return to the matter of how group variability and entitativity affect where group boundaries are drawn. Taking our cue from the Gestalt principles outlined by Campbell, we have already argued that the perception of group entitativity is a complex function of common attributes, common intention/goals/ agency, and group boundaries. More specifically, we have argued that similarity of attributes is a necessary condition for perceiving entitativity and that similarity of goals or intentions— while not a necessary condition for perceiving entitativity—can strongly influence judgments of entitativity (cf. Lickel et al., 2000). Moreover, there are circumstances in which a group can be perceived, explicitly or implicitly, as a natural kind, with concomitant descriptions of the groups “essential nature.” Such groups are perceived as being very homogeneous, with stable defining features, and as having impermeable boundaries (i.e., movement into or out of the category is difficult or impossible). Essence leads to even stronger perceptions of entitativity, but the relation of essence to common goals is complex. Since ethnic or racial groups are among the most likely to be treated as natural kinds, these groups are best characterized as having common attributes but little or no common agency. However, an important consequence of essentialistic attribution is often the attribution of agency or collective intent, so essence seems to be able to create the perception of group agency out of the reality of shared attributes. In short, shared attributes, common goals, and essence all contribute to the perception of entitativity, but not in equal amounts, and not in a simple Guttmann fashion. In the following analysis, we attempt to illustrate the interrelations among homogeneity, agency, essence, and entitativity with specific examples. First, consider groups that are perceived to be both high in homogeneity and in entitativity, such as the KKK, the Nazis, and Orthodox Jews. Consistent with our earlier analysis, the members of such groups are viewed as having many attributes in common, as well as sharing common goals and intentions. Moreover, knowing someone is a member of the group is highly informative with respect to what one can expect that person to be like (high inductive potential). Finally, membership in these groups implies relatively “sticky” characteristics such that the characteristics are unlikely to change even if one moves out of the group. Groups of this sort have very clear, very strong category boundaries, with a tight definition of what the group is like. The category breadth is narrow in that there is a relatively circumscribed pattern of behavior, dress, and mannerisms that qualify one for membership in the group. If an individual does not fall within this highly constrained set of group attributes, he or she will clearly be seen as falling outside the group boundaries. According to our earlier arguments, then, such an individual is likely to be subtyped, to be
74
THE PSYCHOLOGY OF GROUP PERCEPTION
functionally isolated from the rest of the group, and therefore to have little or no influence in modifying the group stereotype. An Orthodox Jew who does not keep kosher, a Nazi who marries a Jew, or a member of the Klan who befriends a black are all engaging in highly disconfirming behaviors. But each is so clearly acting outside the group boundaries that they are unlikely to be represented as members of the group, and their characteristics are very unlikely to generalize to the group as a whole. A second set of examples includes groups whose members are seen as high in homogeneity but low in entitativity, such as feminists, hippies, and fraternity members. The groups in this cell include large social categories with shared attributes that lend a high degree of perceived intragroup similarity, but typically there are at most only broad or general shared goals or intentions. Because group characteristics and group membership are perceived as highly alterable (one can cease holding the views of feminists, acting like hippies, or sharing the values of fraternity members), the group has little sense of essence. In general, the group boundaries are much wider and more permeable than for the first examples, and someone who disconfirms the group tendencies can still fall within the boundaries. Thus, although one’s view of the typical feminist might be a professional middle-aged female, it is certainly possible for an 80-year-old grandmother or a stay-at-home mom to be a feminist. It is even possible for a male CEO of a Fortune 500 company to be a feminist. Knowing about such atypical or disconfirming individuals would lead one to revise the contents of the category “feminists.” If a group of atypical feminists were encountered (e.g., stayathome-moms), one might form a subgroup for these members, but functionally these individuals would still be considered members of the group, a particular “type” of feminists. Clearly it is possible for disconfirmation on certain dimensions to fall outside the group boundaries, as when someone who is a misogynist claims also to be a feminist. In this case, the “feminist” would be subtyped and would not affect one’s conception of feminists. In general, however, because the category boundaries are wider, because there is not the same degree of singularity of purpose, or unalterability of group characteristics, it will be possible for disconfirming information to fall within the acceptable group boundaries and to change one’s conception of the group. A third set of examples includes groups that are neither homogeneous nor high in entitativity. The group members are seen as diverse, bonded together often surreptitiously through a shared attribute, and not much of a group at all. Examples include state workers, college students, and smokers. All three of these groups have some attribute that qualifies individuals as a member of the group. But there is a great deal of diversity among group members and little or no shared purpose or goals. In this case the group boundaries are very wide and so almost anyone, regardless of his or her behavior, can be included within the group so long as the one defining characteristic (state employee, enrolled in college, smoke cigarettes) is present. The stereotype is very weak and can be quite readily be pushed around either by the presentation of new discrepant exemplars or by simply having people bring to mind a different “sort” of exem plar (e.g., Lord & Lepper, 1999). For example, the stereotype of college students might include wealthy, privileged 18- to 23-year-olds. But, a less advantaged, often somewhat older group of students who support themselves and are trying to better their life situations can be brought to mind. In fact, many of the examples in this set are really a collection of subgroups, and one’s conception of the group at a given in time depends on which subgroups are retrieved and considered. In any event, almost all behavior can be included within the boundaries of such groups so long as the defining characteristic (perhaps simply the category label) is present. The final set of possible examples would be high in entitativity and low in homogeneity, which we believe to be either a null set or at least a very sparsely populated set. We have argued that homogeneity of attributes is a necessary, but not sufficient, condition for the perception of high entitativity, and although it is possible to have high homogeneity with low entitativity, it does not make sense, in our view, to have low homogeneity and high entitativity.
THE MENTAL REPRESENTATION OF SOCIAL CATEGORIES
75
Finally, it is important to note that it is possible to shift the entitativity value of all groups through on contextual factors. At times the groups in these cells become organized with respect to some issue. This might be the passage of some public policy that is beneficial to women, or the consideration of a campuswide policy regulating parties at fraternity houses. Consider the low entitative category “state workers,” which includes—among others—university professors, state police, janitors, and neurosurgeons. It is hard to imagine a less entitative group, and yet if there were a public vote on a state retirement plan affecting all state employees, the shared vote could lead both ingroup and outgroup members to view this otherwise heterogeneous category as a strong entity, at least temporarily. Or a proposed citywide ban on smoking in public places might bring together all sort of odd bedfellows. In all of these instances, these groups can temporarily become more entitative because of a newfound focus on some common purpose or goal, and most likely there will be some degree of assumed homogeneity on a few very key attributes as well. It is assumed that when the issue at hand is resolved, however, such groups typically move back to a state of lower entitativity and lower homogeneity. Because of the temporary character of these contextually induced groups, it is unlikely that any would be imbued with essential characteristics. CONCLUSIONS In this chapter we have considered the question of when stereotypic beliefs will be modified by experience with the group’s members. A simple, rationalistic view suggests that stereotypes should be revised when there is a discrepancy between the attributes associated with the group and those associated with its members. One reason this does not happen is because group members who are discrepant from the stereotype are frequently not represented as group members—that is, the poor-fitting group members are functionally isolated from the category (i.e., subtyped). Whether a group member is or is not incorporated within a group’s boundaries depends on several factors, including the homogeneity of a category and the impermeability of its boundaries—factors traditionally associated with the concept of entitativity. It is argued here that entitativity consistents of three components: the degree to which group members are seen as sharing attributes, the strength of agency or perceived commonality of goals and intentions, and the extent to which the group is seen as possessing a strong underlying essence, which in turn is driven by inductive potential and perceived immutability of group characteristics. Although perceived homogeneity is a necessary condition for strong entitativity, it is, by itself, not sufficient. Groups that are high in both entitativity and homogeneity have the tightest boundaries and are therefore the least likely to include poorfitting members, and so are the most resistant to change. Low entitative groups are more likely to allow for change. However, contextual factors by themselves can affect perceived entitativity, apart from any objective characteristics of the group. Thus the language used in discussing the group, ingroup versus outgroup status, and the classificatory level at which the group is considered (which affects ingroup versus outgroup status), can all influence the degree of entitativity, and can allow a motivated perceiver to perceive strong group boundaries even when these might not exist for a more neutral observer (e.g., perceptions of homosexuals by the religious right versus a university student). As such, these contextual factors in and of themselves can prevent belief revision. If the goal is to produce change in a stereotype, then the initial challenge is to somehow break down the entitative structure of the group in order for it to be susceptible to change by disconfirming exemplars. This would mean increasing the perceived diversity of the group (for example, by subgrouping) and by demonstrating the permeability of group boundaries (for example, by exposing the arbitrariness of boundary placement, as well as the possibility of movement into and out of the group). We do not underestimate the difficulty of this challenge.
76
THE PSYCHOLOGY OF GROUP PERCEPTION
NOTES This paper was supported by National Institute of Mental Health Grant MH40662 to the first author and by National Institute of Mental Health Grant MH45049 to the second author. 1. It should be noted that the 149 groups varied on a number of dimensions and an attempt was made to generate groups that would yield all possible combinations of strength and favorability. Nonetheless we did find an overall low negative correlation between strength and favorability (r=−.30). Since we were specifically assessing the strength of group boundaries (ease of entering and exiting a group), we did not include “biological” categories, such as gender, race, or ethnicity.
6 Subjective Essentialism in Action Self-Anchoring and Social Control as Consequences of Fundamental Social Divides VINCENT YZERBYT CLAUDIA ESTRADA OLIVIER CORNEILLE* ELÉONORE SERON* STÉPHANIE DEMOULIN* Catholic University of Louvain at Louvain-la-Neuve *Also at the Belgian National Fund for Scientific Resarch
INTRODUCTION Are you a Muslim or a Christian? Are you a Serb or a Bosnian? Are you a Catholic or a Protestant? Are you a Flemish or a Walloon? Quite often, questions such as these may come across as casual requests aimed at better knowing another person. After all, the various groups to which we belong shape a myriad of characteristics that contribute to making us the persons we are. We can think of such aspects as musical tastes, culinary preferences, or fashion favorites. We may also allude to somewhat more serious features such as political opinions or philosophical references. In most contexts, we are happy to provide the necessary information as knowledge about the likes and dislikes associated to various social groups will likely make the discussion more instructive. In some cases, however, the answer to inquiries of this kind may involve more dramatic costs. Depending on your membership in certain groups, you are likely to be discriminated against at work, denied full access to certain social services, confronted with restrictions on visits to certain countries, and so on. In extreme situations group membership means life or death for the person being asked. Clearly, the world may well be a global village, but cultural divides of all sorts continue to play a most important role in our everyday life. People not only regulate their feelings, beliefs, and behaviors in accordance with their membership in racial, sexual, ethnic, religious, or ideological groups, and the consequences of this “strategy” are far from trivial. In this chapter, we examine some of the consequences of the belief that the members of a group, and members of our own group in particular, share fundamental characteristics that others do not possess. Indeed, we claim that a series of important phenomena emerge upon the evocation of the fact that we and our group members may enjoy the presence of some inherent feature, of an underlying essence, that
78
THE PSYCHOLOGY OF GROUP PERCEPTION
distinguishes us from other groups. In short, we consider the intragroup consequences of the belief that there is a fundamental partition of the world into us and them. The perspective we adopt departs from currently available work on homogeneity, entitativity, and essentialism in several ways. First, our own research on this front has generally examined the causal path flowing from homogeneity through entitativity to essentialism. This is true for our work showing the impact of similarity and homogeneity on the facilitation of dispositional inferences regarding social units (Rogier & Yzerbyt, 1999; Yzerbyt & Rogier, 2001; Yzerbyt, Rogier, & Fiske, 1998) or on the emergence of identification (Castano, this volume; Castano, Yzerbyt, & Bourguignon, in press; for a review, see Yzerbyt, Castano, Leyens, & Paladino, 2000). Other researchers tended to adopt the same perspective to show, for instance, that similarity (Brewer, Weber, & Carini, 1995) or interdependence (L.Gaertner & Schopler, 1998), two factors thought to be antecedents of entitativity, provoked an increase in the level of ingroup favoritism. A similar message can be found in the work by Abelson and colleagues (Abelson, Dasgupta, Park, & Banaji, 1998; Dasgupta, Banaji, & Abelson, 1999) in that greater similarity among group members evokes the presence of shared underlying psychological characteristics. In the present contribution, we review some evidence that entitativity influences subjective essentialism, but we also propose that the reverse causal flow deserves close attention. We thus present new data that speak to the issue of the impact of subjective essentialism on entitativity (Yzerbyt, Corneille & Estrada, 2001). A second distinctive message of the present chapter concerns the specific role of the self. There has been a revival of interest in the role of the self in intergroup processes (Cadinu & Rothbart, 1996; Otten & Wentura, 2001; Sedikides & Brewer, 2001). Indeed, we would like to argue that the self stands out as a key factor in the events that follow the evocation of some shared es sence with fellow ingroup members. Thus, although current wisdom has it that the self very much fades away when social divides become contextually salient (Turner, 1987b), our data suggest instead that the group may well be seen as even more an extension of the self whenever some fundamental characteristic is thought to be shared by all group members. A third important characteristic of this chapter is the variety of dependent variables that are being examined. Indeed, contemporary research on entitativity and essentialism has tended to look at a rather limited range of perceptual and cognitive consequences of the belief in ingroup entitativity and essentialism. In the various studies reviewed below, we intend to show that beliefs regarding the existence of a group essence exert their impact on diverse and yet unexplored dimensions of intragroup relations. In particular, we will concentrate on issues of self-anchoring and social control. Before we turn to the heart of the matter, a first section will provide a reminder of the way we see the various concepts to be used in the rest of the chapter as well as how we conceive of their relationships. In a second section, we will turn our attention to the impact of essentialist beliefs on the emergence of entitativity. In particular, we will show that the conviction that group members share the same underlying properties leads to the exacerbation of the well-known accentuation effect (Tajfel & Wilkes, 1963). Having provided empirical evidence for the idea that people are amazingly skilled at generating homogeneity and similarity whenever essentialist beliefs are present, the two following sections will be devoted to a series of recent studies aimed at showing how subjective essentialism directly affects intragroup processes. Indeed, we will argue that group members holding essentialist beliefs about their ingroup will be very much inclined to rely on self-anchoring. Specifically, in a series of experiments relying on the group polarization paradigm, we will show that even indirect indications of the essential nature of group membership seem to lead people to view their self as a point of reference. In a fourth section, we will present preliminary evidence that people’s beliefs about the nature of group membership also has direct consequences on the reactions that group members manifest toward their fellow group members. A reference to the existence of essential features would seem to alter the way people map
SUBJECTIVE ESSENTIALISM IN ACTION
79
the positions of the different members of the group. This is true for all group members but in particular for the most deviant ones. We will conclude by elaborating a bit on the role of discrimination in triggering the impression that people share the same underlying essence as some others. As we see it, discrimination is a key factor leading victims to construe the social world in discrete categories. By invoking some inherent characteristics that account for the differential treatment they undergo, victims may regain some sense of control over the situation. Hopefully, we will have built a case for the importance of essentialist beliefs in social life. ENTITATIVITY AND ESSENTIALISM As the various chapters of the present volume demonstrate, recent years have witnessed a growing interest in such concepts as entitativity or essentialism. Contemporary research efforts focusing on entitativity, a notion first proposed by D.T.Campbell (1958), started to penetrate social psychology journals and edited books (Abelson, Dasgupta, Park, & Banaji, 1998; Brewer & Harasty, 1996; Dasgupta, Banaji & Abelson, 1999; Hamilton & Sherman, 1996; McGarty, Haslam, Hutchinson, & Grace, 1995; Wilder & Simon, 1998; Yzerbyt, Rogier, & Fiske, 1998) at about the same time that a series of authors, intrigued by Rothbart and Taylor’s (1992) fascinating ideas that some social categories are responded to as if they were natural kinds, presented their initial empirical work on the role of essentialism in social perception (N. Haslam, 1998; Haslam, Rothschild & Ernst, 2000; Yzerbyt, Corneille, & Estrada, 2001; Yzerbyt, Rocher, & Schadron, 1997; see also, Hirschfeld, 1995, 1999). The joint emergence of these concepts is by no means coincidental. Indeed, there are a number of reasons to believe that they are related. Still, we agree with Hamilton et al. (this volume) when they suggest that “we are dealing with quite different and at least somewhat independent concepts.” In other words, entitativity and essentialism should by no means be seen as synonyms. As to the potential confusion resulting from the constant association of these two terms, it may be useful to spell out a few guiding ideas that underlie our research program. When D.T.Campbell (1958) created the neologism entitativity, he defined it as the degree of having the nature of an entity. A more constructive definitional effort came from the Campbell’s provision of a series of cues that he thought was playing a major role in the perception of entitativity. These perceptual cues have since been examined by a variety of researchers (Gaertner & Schopler, 1998; Lickel et al., 2000) and seem to point in two directions. The first covers a series of group characteristics classically associated with work on social categories and stereotyping (e.g., similarity, proximity, group boundaries, etc.). The second set of group features seems more at home in the context of small group research (e.g., coordination in action, common fate, common goals, etc.). Similarly, social psychological work on essentialism has been associated with the close inspection of a series of defining dimensions supposed to be associated with essentialism (N.Haslam, 1998; Rothbart & Taylor, 1992; Yzerbyt, Rocher, & Schadron, 1997). For instance, N.Haslam, Rothschild, and Ernst (2000) asked their participants to evaluate a series of social categories using eight features such as immutability, inherence, and uniformity, among others. These authors identified two dimensions organizing laypeople’s essentialist beliefs. A first dimension corresponded closely to the philosophical concept of natural kind and included beliefs that a social category is natural, has a sharp boundary, immutability, historical invariance, and necessary features. The sec ond dimension, which Haslam et al. (2000) called reification, indicates the extent to which a social category is though to be a meaningful unit with deep commonalties. Loading highly on this factor were beliefs about the uniformity of the category, its inherent basis, and the inductive potential of category membership.
80
THE PSYCHOLOGY OF GROUP PERCEPTION
We applaud Haslam and colleagues’ (2000) endeavor to sort out various features that have generally been associated with essentialism. However, there are also a number of limitations in this work. As a result, we think that finer distinctions within the reification factor may have been obscured. For instance, we suspect that the specific set of characteristics used in their work along with a series of methodological limitations of the study may have contributed to putting the distinction between the two dimensions of natural kind and reification at the forefront. A second problem is that a great deal of research suggests it is quite a different thing to assess the presence of a given characteristic of a target group, say, by asking people to evaluate the overall similarity of the group members and to rely on judgments that are less explicitly related to this same characteristic, by directly requesting evaluations of a series of group members on several personality traits, and only then computing the perceived homogeneity of the group members (for an illustration, see Linville and Fischer, this volume). At a third level, some degree of caution may be needed about respondents’ ability and willingness to express an opinion about sensitive aspects of essentialism regarding a vast series of groups at the same time. In contrast, settings in which perceivers are led to provide rather concrete evaluations about one relevant target group may be more helpful in the study of relationships between various essentialist beliefs. Finally, the contemplation of several characteristics in what appears to be a decontextualized setting tells us little about the dynamic aspects of group perception. Clearly, the impression that people could or could not easily give up their membership in a given group and become member of another group, to take but one example, may be greatly affected by transient structural factors. In order to try and discriminate between the two concepts of entitativity and essentialism (Yzerbyt, Corneille, & Estrada, 2001), we proposed a distinction between the phenotypic and genotypic levels of perception. The phenotype directly refers to the more perceptual aspects generally associated with entitativity. That is, we can talk of entitativity as mainly dealing with surface features of the group members and the group, all aspects that are thought to be easily perceived and judged by an external observer (D.T.Campbell, 1958). In contrast, the reference to genotype allows evoking the more inferential level of social perception, a level usually associated with the idea of essence. To be sure, the word genotype does not mean that we are building on a strict version of biological essentialism, and one can probably get a better grasp of the difference we propose between the genotypic and the phenotypic aspects of group perception by relying on similar distinctions in personality and social psychology. Indeed, when talking about such constructs as traits or attitudes, observers are generally assuming that something deep inside people underlies their behaviors, opinions, and feelings. For instance, when people mention the trait “extroversion” with respect to someone, they use the term to evoke the existence of some disposition that this person has, and they do so by building upon a host of observations. There may well be an ongoing debate and markedly divergent views about the actual causes believed to be responsible for the emergence of a given attitude or a specific trait in any individual, whether learning processes and socialization or genetic factors are invoked, but it remains that constructs such as these capture a level of inference that make them qualitatively different from what is seen as simple indicators. For those students of the role of language in causal attribution and intergroup relations (Maass, 1999; Semin & Fiedler, 1991), a similar distinction is made between the concrete and abstract levels of communication. When communicating about some event by saying that “George is aggressive” rather than “George kicks the dog,” the observer goes beyond the behavioral information and suggests the presence of some “thing” that presumably accounts for what is being observed. In a way, we see a number of advantages in keeping a distinction between a reference to some essence that would not be strictly associated with a pure biological interpretation. Social observers may often react as if they were adopting some sort of “biological framework” when in fact they are referring to influences such as education, religion, culture, climate, and so on. As it turns out, the “inherence” characteristics as it was used by Haslam et al. (2000) is likely to be a proxy for essence even though is was located on the
SUBJECTIVE ESSENTIALISM IN ACTION
81
reification factor. Indeed, it has been found to be moderately correlated with the naturalness factor. Interestingly, our somewhat looser conception of what would be “natural” allows us to understand why people may often be tempted to attribute essential features to artifacts. A popular example is the sharp distinction between the cars made by different national groups. There is undoubtedly something of the Italian artistic touch and sophistication in the Alfa Romeo, and a Mercedes is filled with German strength and hardness. This distinction lies at the heart of what we call subjective essentialism in the social domain (Yzerbyt, Rocher, & Schadron, 1997). When one or several perceptual cues point to the entitativity of a group of people, perceivers are inclined to infer the presence of some essence shared by these people. As a result, they may often end up making strong assumptions about the inductive potential and unalterability associated with group membership. In this sense, the present view on essentialism has much in common with Rothbart & Park’s (this volume) definition of essentialist perception as the ascription to a group of fundamental, basic, or essential dispositions that are regarded as highly enduring and transmitted across time and space through the “mysterious mana” of biology or culture. We would add that the various aspects that D.T.Campbell (1958) saw as cues to entitativity play a major role in triggering such essentialist perception. As a first empirical illustration of this process, our work on social attribution (Rogier & Yzerbyt, 1999; Yzerbyt et al., 1998) aimed at showing that high versus low levels of group entitativity facilitated versus hindered the emergence of essentialist perception. Specifically, we adapted several classic attribution scenarios (E.E.Jones & Harris, 1967; L.D.Ross, Amabile, & Steinmetz, 1977) to make them suitable for a group version and found that people confronted with behaviors emanating from more entitative groups (operationalized by means of group members’ similarity or homogeneity in school affiliation or major) more readily inferred the presence of stable dispositions (for a review, see Yzerbyt & Rogier, 2001). We are not alone in suggesting that the use of traits and adjectives to describe people would be linked to some sort of natural-kind thinking and that people’s tendency to rely on dispositional attribution and so to fall prey to the fundamental attribution error thus stands as an indicator of essentialist perception. For instance, Gelman and Heyman (1999) noted that the use of nouns rather than verbs to describe other people’s behavior leads people to make stronger inferences over time and across settings. The same idea lies at the heart of the linguistic category model (Semin & Fiedler, 1991) and has been nicely illustrated in research on the linguistic intergroup bias (Maass, 1999). Similarity or homogeneity would seem like an obvious way to manipulate the entitativity of the group (Dasgupta, Banaji, & Abelson, 1999). Still, the message put forth by Yzerbyt, Rogier, and Fiske (1998) should certainly not be taken to mean that the presence of shared attributes ought to be equated with entitativity. In fact, we would even argue that homogeneity or similarity is far from being the most powerful cue among the ones listed by D.T.Campbell (1958). Other factors such as the internal organization of the group, the presence of a common fate or the existence of common goals may prove far more efficient cues to signal the entitativity of a group (see Rothbart & Park, this volume). In an attempt to collect evidence for the multiple roots of entitativity and the overall parallelism in their impact, we (Castano, Yzerbyt, & Bourguignon, in press; see Castano, this volume) conducted a series of four studies to show that various aspects of entitativity may impact on people’s reactions to their own group. Distancing ourselves from the literature that say entitative groups come across as being more negative than non-entitative groups (Abelson et al., 1998), we proposed that when it comes to the ingroup, entitativity may actually be seen as a positive feature. One can think of several reasons for this (Castano, Yzerbyt, Paladino, & Sacchi, 2002; Hogg & Mullin, 1999; S.J.Sherman, Hamilton, & Lewis, 1999; Yzerbyt et al., 2000). For instance, group members may see an entitative ingroup as one which is more likely to act upon its agenda and indeed achieve its goals (Yzerbyt et al., 2000). Recent empirical work guided by
82
THE PSYCHOLOGY OF GROUP PERCEPTION
mortality salience induction also suggests that an entitative group is probably more apt to reassure group members when they are reminded about their own physical finitude (Castano et al., 2002). Other authors have also stressed the fact that an entitative ingroup may better fulfill group members’ need for certainty by offering more definite information about the appropriate norms to be adopted (Hogg, this volume; Hogg & Mullin, 1999). Using the European Union as a focal ingroup, our specific prediction was that group members would respond with an increased rather than decreased identification when the entitativity of the European Union happens to be augmented rather than diminished. We expected these findings to obtain mainly, if not only, for participants moderately identified with the European Union. Indeed, more extreme participants should not easily be affected by subtle manipulation of the entitativity of the ingroup, and only moderately identified individuals should possess the necessary latitude to modify their identification. The first study confronted participants with a video of the EU stressing the common fate and common goals of the millions of European citizens in 15 countries. The second study manipulated the similarity aspect of entitativity. Participants were given a list of characteristics of the 15 countries in the European Union and instructed to find differences vs. similarities between the countries and to communicate them to a third party by way of a short text. A third study used a salience manipulation by asking participants to express their opinion of the European Union in an intragroup versus intergroup context. Finally, the fourth study relied on two different levels of boundedness, or, to use Campbell’s words, good continuation, by showing participants that the future borders of the European Union were either not yet entirely decided upon or, instead, very well defined. Here again, participants conveyed their impression about the group in a short written description. In all four instances, the findings confirmed our prediction that the manipulation of a given aspect of entitativity would directly influence the change in identification of moderately identified members. More importantly, because the second and fourth study also secured information about participants’ impressions of the group, we were able to test a mediational model linking our specific manipulations to the participants’ level of identification by means of the degree of entitativity of the European Union as it was perceived by our participants. In both cases, the observed effect of the manipulation on identification (among individuals with a moderate attitude) was fully mediated by the entitativity score given to the texts by blind evaluators. Results such as these go a long way to show that different factors indeed influence the perception of entitativity. Group members were sensitive to the phenotypic information regarding their ingroup, that is, the perception and the representation of the group and group members, in a way that affected their identification with the ingroup. Seen from this perspective, the expression of a strong level of group identification is not unlike a claim that one possesses that which constitutes the essence of the ingroup. In sum, a host of studies seems to suggest that people confronted with a high entitative group are tempted to jump to the conclusion that the members of the group likely possess or probably lack what can be seen as essential char acteristics. Once inferred, these essential, genotypic, features are invoked to account for the observed phenotypic regularities. Along similar lines, a study by Crawford, Sherman & Hamilton (2001) relied on a relearning paradigm to show that, when perceived entitativity of the group is high, observers seemed apt to extract the “essence” of the group and view the individual members as interchangeable parts. Because people who are member of a certain category or a certain group display signs of entitativity, we are inclined to interpret the observed behaviors as expressions of something inherent in the group members, an underlying essence. It is useful to note in passing that we were careful in our studies on the European identification to compare the impact of social categories that varied in their degree of entitativity. Although we would want to argue about the exact meaning some authors give to the notion of aggregate (Gil-White, 2001), there is no ground here for the potential criticism that the comparison involved a social category on the one hand and an aggregate on the other.
SUBJECTIVE ESSENTIALISM IN ACTION
83
All in all, the empirical efforts detailed above are closely related to the observation that social perceivers are Aristotelian rather than Galilean and tend to ascribe the various social phenomena they witness in their daily lives to inherent characteristics of people rather than to the dynamics of the relation and the temporary situational forces. This basic partiality in our perception of the social environment, which some authors call lay dispositionism when applied to individual observers (L.D.Ross & Nisbett, 1991), departs from the Lewinian framework characterizing social psychology and seems to be also operating when the targets of our judgment are social groups. In the next section, we leave the causal path from entitativity to essentialism and turn our attention to the reverse causal flow, asking ourselves how our beliefs about the existence of fundamental characteristics may impact the perception of entitativity. ESSENTIALISM AND THE CREATION OF ENTITATIVITY One of the most fascinating questions in research on entitativity and essentialism is how the perception of groups as being based on essential features affects people’s views regarding these groups. Is it the case, for instance, that groups whose members are thought to share a common essence are also seen as having more surface attributes in common? Is perceived homogeneity or similarity influenced by the belief that group members possess some fundamental characteristics even when the implications of the presence of these underlying features remain entirely unknown? This is the first question to which we devoted our efforts (Yzerbyt, Corneille, & Estrada, 2001; Yzerbyt & Rocher, 2002). To examine this issue, we relied on a paradigm that was adapted from a classic study by Tajfel and Wilkes (1963; Corneille, Klein & Judd, 2002). In an experiment that would turn out to have a lasting impact on social psychology, Tajfel and Wilkes asked their participants to evaluate a set of eight lines. Depending on conditions, each line was presented separately (the no-label condition), the lines were always associated with one of two labels but there was no systematic relation between the length of the line and the specific label (the random-label condition), or the length of the line was systematically linked to the label (the correlated-label condition). The data revealed that a systematicassociation with a trivial category label induced perceivers to accentuate the differences between the two sets of lines. Later work confirmed that there is also a propensity for people to increase the similarity within the categories (see, for instance, Krueger & Clement, 1994; McGarty & Penny, 1988; McGarty & Turner, 1992), at least on dimensions that best allow teasing apart the two categories (Corneille & Judd, 1999). Clearly, the correlated-label condition imagined by Tajfel and Wilkes (1963) should be seen as providing only a minimal set of essentialist beliefs. Consistent with our subjective essentialist perspective, we hypothesized that the imposition of a meaningful as opposed to a trivial category label should invite perceivers to magnify the accentuation effects and to see the groups even more as entities. We thus expected that the availability of a naive theory leading people to assume a fair degree of underlying coherence where little in fact exists would have perceivers strive to reinforce similarity and homogeneity at the surface level. The scenario of the studies was quite simple. We asked students coming from the same population as our participants to write a one-page description of themselves in the context of a social setting. We selected 10 descriptions on the basis of the richness and length, rewrote them, and presented them on separate sheets. In a first study (Yzerbyt & Buidin, 1998, Experiment 1), the 10 descriptions were distributed in two groups of five before being handed out to participants. The participants’ task was to read the first group of five descriptions and form an idea of the first group. They were then to come back to the first description, rate the first person on a series of 10 traits, read the second description and rate the second person on the same 10 traits, and so on. Once the five members of the first group were evaluated, participants were instructed to rate the group as a whole, again using the 10 traits. These steps were then repeated for the second group of
84
THE PSYCHOLOGY OF GROUP PERCEPTION
five members. These traits had been carefully selected so that two scales would include each one of the five dimensions of Norman’s (1963) Big Five factor structure. The analysis of the ratings revealed that 7 of the 10 scales were in fact highly intercorrelated and allowed the computation of a global index of sociability. Apart from the use of one of two random distributions of the 10 descriptions, the other independent variable was the specific rationale given to the participants for the constitution of the two groups. Whereas half of the participants learned that the two groups of five people corresponded to two sets of people waiting for a different bus, the remaining participants were told that the two groups comprised people sharing a different genetic marker. Pretest work had shown that these two kinds of groups were thought to be quite different in their degree of essentialism. Using the ratings given to the group members and to the groups allowed the computing of a variety of indices of central tendency (two group-based evaluations and two target-based evaluations for each participant) and variability (one standard deviation of the sociability index within each group of five targets). Confirming the random nature of the distribution of the 10 descriptions in two groups of five, we found no difference in the group-based or target-based evaluations of the groups. In contrast, and in line with our predictions, the within-group standard deviation was significantly lower when participants were led to think that the groups revolved around different genetic markers than when participants thought that they comprised people waiting for different buses. This finding emphasizes the striking power of beliefs regarding the presence of a higher level of perceived internal coherence in the groups. A second study (Yzerbyt & Buidin, 1998, Experiment 2) relied on the same scenario with the difference that the descriptions were first ranked on the basis of the mean evaluations given in the bus condition of the random study. In other words, one group comprised the five most sociable targets and the other was made of the five least sociable targets. Within each group, the five descriptions were then presented in a random order. Aside from the fact that participants were confronted with one of two orders of presentation, depending on whether the more social group was presented first or second, the crucial experimental manipulation concerned the alleged nature of the groups. In the bus condition, the two groups were said to comprise people waiting for a different bus. Participants in the gene condition learned that the two groups consisted of people with different genetic markers. As expected, the nature of the beliefs influenced the perceived entitativity of the groups. Participants rated the sociability of the five group members to be more similar when they were informed that the group members shared the same genetic marker than when they thought that the groups comprised people waiting for a different bus. A third condition, in which participants were simply informed that they would be confronted with descriptions of 10 people, was also added in the study. Because there was no mention of the existence of two groups in this condition, participants were not asked to rate the groups in and of themselves. Supporting our idea that the provision of a meaningful basis of categorization facilitates the emergence of accentuation, the built-in difference between the more sociable and the less sociable group was qualified by the a priori beliefs about the group. Participants saw little difference between the sociability of the two groups when the various descriptions allegedly originated from 10 people or when the two groups were said to comprise five people waiting for a different bus. In sharp contrast, the mention of some essential feature associated with group membership led to the perception of a significantly stronger difference between the two groups. Other empirical demonstrations of perceivers’ enormous potential for accentuation when groups are organized around essential features can be found in a study using the illusory correlation paradigm (Yzerbyt, Rocher, McGarty & Haslam, 1997; see Yzerbyt & Rocher, 2002). The illusory correlation paradigm has become a classic and there is thus no need to dwell on its detailed aspects here (Hamilton & Gifford, 1976). Suffice it to say that participants were presented with information about the behaviors of the members of two groups, a majority and a minority. Most behaviors are positive but the proportion of
SUBJECTIVE ESSENTIALISM IN ACTION
85
positive to negative behaviors is actually the same in both groups. A classic finding is that people tend to associate more negativity with the minority group. Over the years, it has become clear that the prevailing explanation is one in terms of differentiated meaning (Berndsen, Spears, McGarty, & van der Pligt, 1998; McGarty, Haslam, Turner, & Oakes, 1993; McGarty, Yzerbyt, & Spears, 2002; for a review, see Berndsen, Spears, Van der Pligt, & McCarty, 2002). Various features of the situation induce perceivers to differentiate the two groups and, because there seems to be a subtle advantage for the initial working hypothesis that the largest group is positive, the final impression is one of more negativity among members of the minority. To the extent that the differentiated-meaning explanation rests on the idea that participants are likely to confront a working hypothesis against the data, we (Yzerbyt, Rocher, McGarty & Haslam, 1997) reasoned that one key factor that may affect the subjective validity of this hypothesis could be the belief that group membership indeed rests on the presence of some fundamental feature. As expected, compared to control participants who received no other indication about the groups than that they were real groups out there, participants informed that the groups had been constituted by clinical psychologists, a profession thought to know a great deal about essential features of human beings, ended up with a much stronger illusory correlation. In contrast, participants told that the people had been assigned to a particular group by means of a computer program, a method taken to show only limited consideration for people’s deep underlying characteristics, failed to show any sign of illusory correlation. Like the above series of experiments, this study provides compelling evidence that perceivers’ naive theories about the underlying essence of a group affect their construal of the group as a social entity. When participants were informed that groups consisted of people sharing some inherent property, they created similarity and homogeneity at the surface level. Data such as these stress the role of essentialism in intergroup relations because the nature of the beliefs causes a perception of similarity among group members. In the following section, we examine the impact of essentialist beliefs on the way people react to fellow ingroup members. ESSENTIALISM AND GROUP POLARIZATION As members of groups, we form beliefs about those very social entities to which we belong. Is it the case that insiders are sensitive to the assumed nature of their group and shape their behaviors accordingly? To begin to address this issue, we set out to examine the impact of naive theories about the origin of the ingroup on group decision-making processes. In a series of studies (Yzerbyt & Estrada, 1999), we examined the role of subjective essentialism in the construction of ingroup consensus by relying on a standard grouppolarization paradigm. Participants were asked to volunteer for a study on opinion and decision making and were presented with a series of 10 risky dilemmas taken from the group polarization literature. Typically, such dilemmas describe a situation in which a target person is confronted with one line of action that is unattractive but involves no risk at all and another line of action that is much more attractive but entails some level of risk. The task of the participants is to read each dilemma and indicate the highest level of risk for them to continue to encourage the target person to select the attractive line of action. The study took place in two sessions separated by one week. In a first session, up to 5 participants were first given a trial dilemma to familiarize them with the dependent measures. They were then given a series of 10 dilemmas and ample time to come up with their answers. Upon completion of the dilemma questionnaire, participants were asked whether they would agree to come back to the laboratory along with four other participants in order to study group decision making. In addition to setting an appointment time for the second session, participants were invited to take part to a test allegedly aimed at identifying the presence of certain personal characteristics. Specifically, they were asked to put a sample of their saliva on a small
86
THE PSYCHOLOGY OF GROUP PERCEPTION
colored test paper. The female experimenter explained the result of the test could or could not be important for the second part of the study and that participants would receive more information if necessary. She then displayed the five colors that could result from the saliva test and explained that each color was associated with a different set of objective characteristics and that a few minutes were needed for the test paper to take on a particular color. Taking advantage of participants examining the names given to the various colors, the experimenter randomly selected one of five precolored test papers and presented it as being the result of the participant’s saliva test. Participants were then informed of the result of the saliva test that characterized them and asked to fill in the record sheet accordingly. An appointment was made for a second session. In the second session, participants (from 3 to 5) were seated around a table in the middle of a large room. Control participants were informed that they would be presented with the same dilemmas as the ones used during the first session and that they were expected to reach a consensus in their answers using a maximum of 60 minutes for the 10 dilemmas. Experimental participants were also asked to reexamine their record sheet and to notice that they were all associated with the same color, hence with the same set of objective characteristics. Actually, the assignment to the conditions was purely random; the only constraint was that participants were assigned to groups by test color. At the end of the time given for the group task, participants were asked to individually sit either at one of the four tables placed against the walls of the lab room or at the central table. They were invited to go over each dilemma anew and to indicate what they now thought was the appropriate answer for each problem. Finally, participants were given a series of manipulation checks. They were then debriefed, thanked, and dismissed. Manipulation checks confirmed that our manipulation of the naive theories was successful. In line with the standard procedure used in group polarization research, we first looked at the answers provided during the initial, individual phase. Not surprisingly, participants later assigned in one of the two conditions initially expressed similar views on the dilemmas. As a means of evaluating the presence of group polarization, we then compared the group answers to the initial individual answers. The average reactions to the 10 dilemmas revealed the presence of a significant risky shift. We had no clear expectations regarding the amount of polarization that could emerge in each one of the two conditions. On the one hand, it is possible to argue that people who are told that they share the same essence would be less ready to change their initial position. This could be because members of essentialized groups would already see themselves as a prototypical instance of the group or, alternatively, because they would like to retain their individual viewpoint as much as possible in order to affirm their distinctiveness (Brewer, 1991). On the other hand, the information regarding the underlying essence could facilitate the operation of informational and normative influence among the members of the group and result in greater polarization. Although the means were more consistent with the first than with the second conjecture, an ANOVA analysis indicated that the total amount of polarization on the 10 dilemmas was comparable in these two conditions. This pattern of findings means that all participants manifested polarization when confronted with a group decision. Because we expected that naive theories about the foundation of the group would indeed influence the way people react toward other group members’ positions, we compared how participants in the different conditions changed their individual decision into a group decision. To do this, we computed the squared difference between the initial individual answers and the group answer. The data revealed that the average amount of individual change was lower in the experimental than in the control condition. Such a pattern suggests that members of essentialized groups felt less of a need than control participants to abandon their individual position. To evaluate this interpretation more directly, we also computed the squared difference between the initial and the final individual answers. As expected, the data showed that the average impact of the group discussion was more limited when participants thought that all group members shared the same
SUBJECTIVE ESSENTIALISM IN ACTION
87
essence. In sum, essentialized participants were more likely to maintain their initial position despite the experience of the group discussion. These data provide intriguing evidence that people do not react in the same way when they think of themselves as a member of an ad hoc group or as a member of a group that is based on the shared presence of some deep characteristics. Although we provided our experimental participants with minimal information about the essential nature of the groups, such knowledge had a definite impact on the way they reacted to the situation. In the present case, the simple activation of naive theories reinforcing the ontological status of the group changed the way the members of the groups reached consensus in a decision-making context by minimizing their willingness to modify their initial position. In other words, the shifts toward the “new” consensual answer appeared to be smaller and more evenly distributed among members of essentialized than nonessentialized groups. One possible explanation for this finding is that the members of essentialized groups saw the positions held by the different members of their group as being less different from each other. In a related vein, the reference to a strong ontological status of the group may have encouraged these participants to consider themselves as a prototypical incarnation of their group. As a result, members of essentialized groups may feel less of a need to abandon their individual positions. As we indicated, one intriguing alternative interpretation of our findings is that members of essentialist groups were more strongly opposed to the group decision and more reluctant to give up their distinctive features than members of nonessentialist groups. In this perspective, essentialized participants stick to their personal views in order to reaffirm their sense of uniqueness (Brewer, 1991). ESSENTIALISM AND SELF-ANCHORING Obviously, the former explanation is consistent with the view that people rely on their beliefs regarding the underlying essence of their group to create similarity. Along the lines developed by Cadinu and Rothbart (1996) in their work on self-anchoring (see also Otten & Wentura, 2001), group members would then see the rest of their group as less different from themselves. That is, they build upon their own characteristics and start seeing the group as a more coherent entity. This phenomenon bears some resemblance to social projection (Krueger, 1998). In contrast, the latter explanation stresses the fact that deep sharedness (especially when it concerns groups that have not been freely chosen) can be experienced as oppressive and lead group members to reaffirm their idiosyncrasy. In terms of Brewer (1991) and her theory of optimal distinc tiveness, one would thus say that group members would be reluctant to adopt other people’s views when their need for differentiation gets activated by the explicit reference to a shared underlying characteristic. In an attempt to replicate the above findings and to collect empirical evidence that would allow us to begin disentangling these two explanations, we designed a second study (Yzerbyt & Estrada, 1999, Experiment 2). In order to simplify the logistics of the experiment, we developed a scenario in which participants were confronted with a virtual group rather than with actual participants. Concretely, participants were called into the laboratory and informed that the study concerned the impact of various perceptual types on decision making. As a means of securing an initial attitudinal measure, participants were given an answer sheet and asked to provide their attitude regarding a controversial issue, namely, whether students reaching the end of the first level of their curriculum in psychology should be forced to undergo a psychological checkup. Participants indicated their reactions on a scale from 1 (totally against) to 7 (totally in favor) and wrote the various arguments in support of their position. Next, participants were presented with two filler tasks aimed at reinforcing the cover story. At this point, experimental participants were invited to take a perceptual test allegedly aimed at uncovering the presence of certain deep personality
88
THE PSYCHOLOGY OF GROUP PERCEPTION
features. This test was made to comprise 7 groups of 3 geometrical figures, and the participant’s task was simply to select the most preferred figure of the three. At the end of the test the experimenter explained that, although the test had given reliable information about the perceptual type of the participant, she wanted to further ascertain this result by means of a well-established saliva test. In all cases, the saliva test was said to have confirmed the perceptual type of the participant. All participants were then given a chance to examine the answers to the controversial issue (both the scale values and the arguments) of four other participants that, they learned, had already taken part to the experiment. Experimental participants were informed that all four students had the same perceptual type as themselves. The opinions distributed to the participants were actual opinions randomly selected from a pool of 9 original answers. Each random selection was yoked across control and experimental participants. Participants were asked to imagine a two-hour group discussion with these four people and presented with a series of questions concerning their likely opinion about the controversial issue at the end of the two-hour discussion, the group’s final opinion, the most representative member of the group, their perception of homogeneity and entitativity of the group at the beginning and at the end of the discussion, the prototypicality of each one of the five members of the group, and the degree of affinity and similarity between themselves and the group. If the pattern observed in our first study is mainly the consequence of self-anchoring and social projection, the provision of naive theories about the essential nature of the group should lead participants in the present experiment to see themselves as more typical representatives of the group when they have been informed that they share a fundamental characteristic with the other members of the virtual discussion group. In sharp contrast, if some kind of search for distinctiveness underlies the inertia in individual positions evidenced in the first study, we should rather observe a general reluctance in presenting oneself as a typical representative of the group. A similar rationale holds for the remaining dependent measures. Clearly, the observed data provide strong support for our self-anchoring hypothesis and allow us to question the viability of an optimal distinctiveness interpretation of the data. Interestingly, we observed no difference in the individual positions collected either before or after the provision of the opinions of the four virtual participants depending on the experimental condition. This finding notwithstanding, the withincondition correlations between the final individual position and the estimated final group position indicated that there was an impact of the group opinions on the individual position only when the group was said to be based on the shared existence of a perceptual type. We also computed the absolute distance between the group’s opinion and the final individual answer. Here again, the individual was significantly closer to the group’s opinion when the experimenter had referred to the essence of the group. The distance between participants’ final individual position and the position of the most typical group member was also significantly smaller when essential attributes of the group had been evoked. The image of the group was also significantly affected by the instructions. Compared to control participants, experimental participants saw the group as significantly more homogeneous and entitative. Interestingly, this was true only on the questions pertaining to the end of the discussion. In other words, compared to control participants, experimental participants did not think that the group was more entitative at the beginning of the discussion. What did participants think concerning the prototypicality of themselves and their group members? On average, the five members of the group were seen as more representative of the group when the experimenter had mentioned the presence of some essential group characteristic. More importantly, the participants assigned to themselves a much higher level of prototypicality in the essentialized groups than in the control groups. Finally, experimental participants saw much more affinity and similarity between themselves and the rest of the group than control participants.
SUBJECTIVE ESSENTIALISM IN ACTION
89
Clearly, although further research is certainly needed to disentangle the interpretation based on selfanchoring and social projection (Cadinu & Rothbart, 1996; Krueger, 1998; Yzerbyt, Corneille, Estrada, 2001) from the one derived from optimal distinctiveness (Brewer, 1991), data such as these provide compelling evidence that naive beliefs about the nature of groups may indeed have a substantial impact on group decision making. The present line of research also proves interesting in the context of current work on intragroup dynamics (Abrams, Marques, Bown, & Henson, 2000; Marques, Abrams, Paez, & MartinezTaboada, 2000; Marques & Yzerbyt, 1988; Marques, Yzerbyt, & Leyens, 1988; Coull, Yzerbyt, Castano, Paladino, & Leemans, 2001; for reviews, see Marques, Abrams, Paez, & Hogg, 2000; Yzerbyt, Castano, Leyens, & Paladino, 2000). As a matter of fact, it would seem that the way people react toward deviant ingroup members may be influenced by their beliefs about the nature of what brings the group members together in the first place. This is the focus of the next section. ESSENTIALISM AND REACTIONS TO DEVIANCE Group life is far from always being harmonious. Dissension and consensus are at the heart of group dynamics. How people react to the presence of a deviant group member is of utmost importance in the evolution of a group (Asch, 1956b; Moscovici, 1976; Schachter, 1951). Often, some form of negotiation is being initiated. Either marginal views are set aside to the full benefit of the dominant wisdom or the challenging perspective gets somehow integrated into the most popular position and new convictions emerge from the process (for review, see Van Avermaet, 1996). When the gap between the deviant and the rest of the group is too large, the reaction of the rest of the group is likely to be extreme, and schisms may even take place (Sani & Reicher, 2000). That is, the deviant is derogated (Marques, Yzerbyt, & Leyens, 1988) if not excommunicated (Castano, Paladino, Coull, & Yzerbyt, in press; for reviews, see Marques, Abrams, Paez, & Hogg, 2000; Yzerbyt, Castano, Leyens, & Paladino, 2000). On the basis of Yzerbyt and Estrada’s (1999) findings, we reasoned that, compared to members of nonessentialized groups, members of essentialized groups may indeed be inclined to react differently to deviant group members than members of less essentialized groups. Although they would generally tend to see group members in a more homogeneous way, i.e., a tendency corresponding to some sort of assimilation, they would also see a strongly divergent position as more irreconcilable with the dominant perspective, i.e., a pattern akin to a contrast effect. In sum, we expected members of essentialized groups to be more eager to affirm the homogeneity of the group and to exacerbate the reaction toward the deviant when its position is too distant from the dominant norm. The manifestation of strong opposition to the norm of the group would be taken very seriously and provoke a series of reactions such that reconciliation would seem difficult to achieve. The study we designed to test this hypothesis (Estrada & Yzerbyt, 2002) once again had participants take part in a virtual discussion group. First, participants were given an answer sheet and asked to communicate their attitude toward the same controversial proposal as the one used by Yzerbyt and Estrada (2000, Experiment 2), that is, students reaching the end of the first level of their curriculum in psychology should be forced to undergo a psychological checkup. Participants indicated their answer on a scale from 1 (totally against) to 7 (totally in favor) and provided various arguments in support of their attitudinal position, which they wrote on the answer sheet. As a means to later create the impression that they shared some underlying feature with the other members of their discussion group, experimental participants were invited to take a perceptual test allegedly aimed at uncovering the presence of certain deep personality features. The same procedure was used as in Yzerbyt and Estrada (1999, Experiment 2).
90
THE PSYCHOLOGY OF GROUP PERCEPTION
All participants were then given a chance to examine the scale values and the list of arguments provided by four other participants that, they learned, had already taken part in the experiment. Experimental participants were informed that all four students shared the same perceptual type as themselves. All participants then received a series of three pro-norm opinions, that is, opinions that were on the same side of the scale as the one expressed by the participants themselves, be it in favor or against the proposal. The nature of the fourth opinion distributed to the participants allowed manipulating the extremity of the deviant opinion. Whereas half the participants were provided with a moderately anti-norm opinion, the other half read a radical anti-norm opinion. It is important to note that the deviant was always holding an opinion which was located across the midpoint of the attitudinal scale, i.e., an anti-norm position. Participants were asked to imagine a two-hour group discussion with these four people and presented with a series of questions concerning the four group members’ opinions about the controversial issue, their own position at the end of the two-hour discussion, the group’s final opinion, their perception of groupness at the end of the discussion, the prototypicality of and their impression of each one of the five members of the group, the identity of the most and least representative member of the group (the latter measure allowing verification that participants had taken good notice of the anti-norm deviant), the distance between the participant and the least representative member of the group, and the degree of affinity and similarity between themselves and the group. Results indicated that the initial attitudes of the participants were the same in the various cells of the design. Also, and importantly, participants’ initial agreement or disagreement with the proposal did not influence the pattern of findings. As a result, all analyses were conducted on the aggregated set of data. Confirming the success of our manipulation, the radical anti-norm deviant was seen as significantly less representative of the group’s opinion than the moderately anti-norm deviant. Also, participants indicated that they saw more difference between themselves and the radical anti-norm deviant than between themselves and the moderately anti-norm deviant. Interestingly, participants’ perceived difference between themselves and the deviant was more pronounced when the group was essentialized. At a more global level, the perceived affinity and similarity among group members and the perceived distance between participants’ initial position and the final group position confirmed that essentialism had the predicted effect. Still, the direct measures of groupness did not reveal the presence of different impressions among participants in the various conditions. Several explanations may account for this somewhat unexpected finding, the most reasonable being that the two measures of groupness may not be ideal in the present case. Indeed, we relied on the same measures we used before in settings where people were confronted only with pro-norm group members. Although these measures proved sensitive enough when all participants shared largely similar views about the proposal, they may actually be too crude to capture the subtle differences between the four conditions created here. The key measures concerned participants’ indication of the likely position of themselves and each one of the group members at the end of the discussion. We computed three different indices to examine this issue. One index corresponded to the average squared distance between the deviant and the three pro-norm members of the group, another to the average squared distance between the deviant and the participants themselves, and a last one to the average squared distance between the participants themselves and the three pro-norm members of the group. Not surprisingly, the more extreme the deviant was, the more participants indicated that they perceived a distance between the deviant and the three other group members. Also, the more extreme the deviant was, the more distant the participants saw themselves from the target. Although the interaction failed to reach a conventional level of significance, the pattern of means suggested that the distance between participants and the radical deviant was perceived to be greater when the group had been essentialized than when it was not
SUBJECTIVE ESSENTIALISM IN ACTION
91
essentialized. Finally, and confirming this trend, a very significant interaction was found with respect to the squared distance between participants themselves and the three pro-norm members of the group. Specifically, the proximity among the four pro-norm individuals was perceived to be strongest in the essentialized groups confronted with a radical deviant and lowest in the nonessentialized groups confronted with a radical deviant. In sum, this study provides promising evidence of the potential impact of essentialist beliefs on people’s reactions to deviance. On the one hand, we replicated our earlier finding that essentialism goes hand in hand with a stronger propensity to manifest self-anchoring, as indicated by the lower squared distance between the initial attitude and the final position of the group. On the other, we found that essentialist beliefs regarding one’s group triggered stronger reactions toward an extreme deviant, potentially precluding compromise and rehabilitation. Reactions such as the ones observed in the above study are reminiscent of a number of responses of caution manifested toward ingroup members or candidates for the ingroup (for a review, see Yzerbyt et al., 2000). A case in point is Brewer’s (2000) insightful analysis of the various factors that may limit the effectiveness of crossed categorization policies in promoting smoother contact between members of different subgroups. She notes that when people simultaneously belong to two different categories, they may embrace an exclusive versus inclusive path. Whereas the latter strategy would have them consider any person who shares at least one category membership as an ingroup member, the former pushes people toward seeing targets who possess both memberships as their sole ingroup fellows. In the present context, we would see the essentialist stance as one that promotes an exclusive attitude with respect to group membership. We suspect that an essentialist posture would thus also lead to stronger manifestations of ingroup overexclusion effect (Castano, Yzerbyt, Bourguignon, & Seron, 2002; Leyens & Yzerbyt, 1992; Yzerbyt, Leyens, & Bellour, 1995). Clearly, there is growing evidence for the fact that group members holding essentialist beliefs about group membership are prompt to see homogeneity among group members, to use their self as an anchor to evaluate ingroup members, and to express less tolerance toward clear instances of deviance. In our conclusion section, we would like to sum up the findings and offer some speculative thoughts about the factors that we believe may foster the emergence of essentialist beliefs in the first place. CONCLUSION In the present chapter, we have been stressing the benefits of examining the relations between entitativity and essentialism more closely. Complementing earlier work showing the impact of cues of entitativity on the emergence of essentialism, we have been pursuing the idea that essentialist beliefs may indeed have a series of noticeable effects on the perception of groups and group members. Our specific focus was on the consequences of subjective essentialism regarding the ingroup on the way people react toward the other group members. A consistent finding was that participants were much more similar to each other in their convergence toward a group position when they believed in the essential nature of their group. We argued that this was due to people’s tendency to see themselves as full-fledged representatives of the group whenever they are convinced they share some deep underlying feature with other group members. In contrast to what would be predicted on the basis of optimal distinctiveness theory, we did not find that essentialized participants were keen to distance themselves from the rest of the group. Clearly, intragroup differentiation was not the main motive underlying the general inertia in people’s positions. In all likelihood, the compromise within the group was obtained on the basis of the individual members’ conviction that they were prototypical instances of the group. This tendency to rely on self-anchoring when one group gets
92
THE PSYCHOLOGY OF GROUP PERCEPTION
essentialized should probably be moderated by a series of factors yet to be explored. Indeed, it is possible that the members of an essentialized group would be happy to differentiate themselves from others on other dimensions than the one used for the study. In any event, further research should help us clarify this issue. Also, it is likely that some manifestation of increased adoption of group characteristics would take place when the normative dimensions are well specified. A second finding concerns the impact of the essential definition of the ingroup on people’s reactions to deviance. As it turns out, the pattern of data indicates that members of essentialized groups tend to see the members of their group as forming more of a homogeneous block in their attitudes and opinions about a key issue. This is not unlike what was found a long time ago by Sherif and Hovland (1961) in the attitudinal domain. Indeed, these authors observed that the way individuals react to a series of attitudinal statements was largely a function of their ego-involvement. When people’s had a vested interest in the particular issue, the number of categories in which statements were distributed was reduced substantially. Moreover, the number of attitudinal statements deemed neither acceptable nor unacceptable decreased, and the two extreme categories of acceptable and unacceptable statement became more populated. This response is not unlike the reaction of members of essentialized groups toward the radical deviant as well as the pro-norm members. One distinct feature of the empirical work presented in previous sections concerns the fact that we somehow informed experimental participants that the targets of judgment or themselves shared some underlying feature with their fellow group members. That such a straightforward instruction leads outside observers and members of essentialized groups alike to boost the level of entitativity of the target group and the homogeneity of its members is no trivial finding. Still, real life settings may tend to differ from the lab environment in precisely that aspect. That is, although we can think of some cases in which people are being explicitly told about the presence of deep underlying features, it is doubtful that such information is always made available. It would therefore seem important to be able to point out the various factors contributing to the emergence of essential beliefs in front of one or several target groups. A good place to start our analysis is to come back to the various cues of entitativity and to their role in suggesting the presence of some underlying essence among group members. Chief among these were similarity, shared goals, and group boundaries. Building upon our own findings and those of several other researchers, we would argue that social conflict in general and discrimination in particular should play a pivotal role in the emergence and perpetuation of subjective essentialism. Indeed, discrimination would be the factor par excellence that establishes deep sharedness among ingroup members because it involves many of the entitativity cues identified by Campbell, such as similarity, common fate, and impermeability. For instance, research shows that discrimination has the capacity to encourage group members to accentuate ingroup homogeneity (Jetten, Branscombe, Schmitt, & Spears, 2001; Rothgerber, 1997; Schmitt & Branscombe, 2001). Building upon this first set of findings, we would predict that discrimination likely invites group members to rely on an essentialist appraisal of the situation (C.L.Martin & Parker, 1995). As can be seen, a basic tenet of our subjective essentialist view (Yzerbyt, Rocher, & Schadron, 1997) is that the ascription of essence to a social group is not only a continuous, rather than an all-or-none phenomenon, but also highly sensitive to the structure of the group relations rather than static and impervious to drastic changes in such aspects as relative status and hierarchy. As far as the first aspect is concerned, we would like to argue that essentialist beliefs indeed come in various levels of intensity. The data from Haslam and colleagues (1998) offer convincing evidence that widely different degrees of essence can be attached to different social groups and that is not like a social group is either fully or not at all essentialized. As for the second characteristic, we propose that subjective essentialism is a dynamic reaction that is most sensitive to the features of the social context and the peculiarities of social perceivers. For some people and in some circumstances, the confrontation with a given group of people may not give way at all to
SUBJECTIVE ESSENTIALISM IN ACTION
93
essentialist beliefs. For other people or in a somewhat different setting, the same group of people may insinuate the idea of an underlying essence. Examples abound showing that the social context may in fact modulate the extent to which certain categories of people get essentialized. Even people working in a firm may come to be seen, by others and even more so by themselves, in a manner that is strongly reminiscent of an essentialist stance. In Belgium, the 70-year-old national air company, Sabena, recently met with a traumatic bankruptcy. This social tragedy led everybody in the country to start talking of “Sabenians.” Belgian citizens and company workers alike were suddenly made aware that a great many Sabenians had somehow been living in a separate world. Due to specific job constraints, pilots, crew members, and even ground workers were hardly meeting anybody else outside of their work context, most of them finding a spouse or a partner within the company. More surprising, the number of employees who actually had a parent who had worked for the same company was astoundingly high. There is now an official website specially reserved for Sabenians where members of the company can find help and comfort from former colleagues, as it would seem that only congeners truly understand each other. One pilot recently committed suicide and, more than six months after the official end of corporated activities, several hundred Sabenians attended the funeral wearing the uniform of the company. We suspect that many of these workers will take quite some time to alter their way of thinking about themselves, and how many of them will be able to embrace new identities remains an open question. In sum, there is little doubt that most Belgians came close to having essentialized Sabenians, much as the company employees seem to have done themselves. As the above situations illustrate, the study of subjective essentialism with respect to social groups is profitably distinguished from the strict biological interpretation of the concept sometimes promoted in anthropological work on essentialism (Gil-White, 2001). The present efforts are concerned with a much larger question than the issue of whether social observers are entitled to treat a particular group in terms of species or not. Rather, the aim is to gain a better understanding of the multifaceted consequences of endowing a social group with some essence as well as the conditions that lead to such a conception. The focus is thus on groups as perceived by the observers, along with the cognitive and social implications of adopting an essentialist approach. To delineate once more our position regarding subjective essentialism in the social domain, we see subjective essentialism about social categories as the conviction that what people do, think, and feel is the expression of their deep nature (Rothbart & Park, this volume). In other words, observed behaviors are taken to derive from something that is largely shared by other group members and is endowed with a substantial level of inertia, be it associated with biology, culture, early socialization, or any other factor (for closely related views, Levy, Stroessner, & Dweck, 1998; Plaks et al., this volume). The belief that some kind of essence plays a causal role in shaping people’s behaviors, opinions, and feelings is at odds with the idea that situational constraints are the main culprit in what takes place in the social arena, which can change both dramatically and rapidly. Given the close connection with the idea of dispositionism in interpersonal perception (L.D.Ross & Nisbett, 1991), the tendency to infer the presence of deep inherent features among social targets can also be seen as a process of “social attribution” (Yzerbyt & Rogier, 2001). Clearly, the data collected in our studies indicate that there is a series of nontrivial phenomena that derive from the apparently innocent belief in the presence of some essence shared by the group members. We are only beginning to see the potential lessons of the present series of studies for our knowledge about group and intergroup behavior. A full specification of the impact of essentialist views in the perception of social groups awaits further investigation. Also, we are beginning to orient our work toward the identification of the various factors that generate such views in the first place. Hopefully, the provision of a global picture
94
THE PSYCHOLOGY OF GROUP PERCEPTION
regarding both the antecedents and consequences of subjective essentialism should give access to some of the key dimensions of social conflict and help us find a way toward their resolution.
Section 2 Antecedents and Consequences
7 In the Eye of the Beholder Lay Theories and the Perception of Group Entitativity, Variability, and Essence JASON E.PLAKS University of Washington SHERI R.LEVY State University of New York at Stony Brook CAROL S.DWECK Columbia University STEVEN J.STROESSNER Barnard College—Columbia University
Although the last half-century has seen widespread efforts to improve intergroup relations, it is clear that stereotypes continue to play a fundamental role in everyday social thought (e.g., Dovidio & Gaertner, 2000; Schuman, Steeh, Bobo, &: Krysan, 1997). Two central elements of stereotyping are expectations of group entitativity (viewing a collection of persons as a single, coherent entity) and group homogeneity (viewing a group as composed of highly similar and interchangeable members; G.Allport, 1954; Brewer & Harasty, 1996; Hamilton & Sherman, 1996; Park & Judd, 1990). Though there are important distinctions between entitativity and homogeneity (Hamilton & Sherman, this volume; Haslam, Rothschild, & Ernst, this volume), there is mounting evidence that when either of these assumptions are in place, perceivers tend to feel it is fair game to extend sweeping characterizations to an entire group and to judge an individual member solely by his or her group membership (Brewer & Harasty, 1996; Gaertner & Schopler, 1998; Hamilton & Sherman, this volume; Park & Judd, 1990; Ryan, Bogart, & Vender, 2000; Yzerbyt, Corneille, & Estrada, 2001; Yzerbyt, Rogier, & Fiske, 1998). Thus, the degree to which people do or do not harbor expectations about group coherence and homogeneity is of great theoretical and practical interest. What leads people to expect high group entitativity and high group homogeneity? In response to this question, a number of researchers have followed D.T.Campbell’s (1958) Gestalt-inspired approach and focused on the inherent qualities of groups (e.g., size, duration, shared fate) that may influence how they are perceived (e.g., Brewer & Harasty, 1996; Lickel et al., 2000; Linville, Fischer, & Salovey, 1989; McConnell, Sherman, & Hamilton, 1997; McGarty et al., 1995; Park & Judd, 1990; Simon, 1992a). For example, recent research has isolated, through correlational analyses, several group properties that are strongly related to perceptions of entitativity (Lickel et al., 2000). In a similar vein, research on the outgroup homogeneity effect (perceiving members of an outgroup as more similar to each other than members of an ingroup), although stemming from a different research tradition, has also on occasion investigated how
IN THE EYE OF THE BEHOLDER
97
group features such as size and majority or minority status influence observers’ perceptions of group homogeneity (e.g., Bartsch & Judd, 1993; Linville, Fischer, & Salovey, 1989; Park & Judd, 1990; Simon, 1992a; but see Brauer, 2001). PERCEIVERS’ LAY THEORIES: THE OTHER END OF THE EQUATION While it is clear that this target-centered approach has yielded important insights into how groups are mentally represented, in this chapter we describe how the literature can be broadened and enriched by a parallel, more direct focus on variability at the perceiver end of the equation (see also Guinote, this volume; Hewstone, this volume; McConnell, 2001; Yzerbyt, Corneille, & Estrada, this volume). We propose that group perception is significantly in the eye of the beholder; that is, different a priori beliefs about human nature can guide people’s perception of both group entitativity and group homogeneity. This perspective, it should be noted, does not stand in opposition to existing views of entitativity and homogeneity. Rather, we suggest that an understanding of the systematic influence of perceivers’ background theories complements the existing research in at least three important ways. First, rather than assuming that all people reach relatively uniform perceptions of the homogeneity and entitativity of a given group, we propose that understanding perceivers’ lay theories can open a window to systematic variability in such perceptions. That is, knowing perceivers’ lay theories may foster a better under standing of who is more likely to see a collection of people as a monolithic group (and in turn, who is more likely to stereotype). Second, isolating such differences may promote a fuller understanding of the processing mechanics of group perception. Perceivers may have at their disposal multiple options for how to interpret, represent, and elaborate incoming group information; which option is followed may depend importantly on the operating assumptions that perceivers are working with. Third, to the degree that these operating assumptions can be altered, a focus on their impact might suggest concrete strategies for reducing stereotyping and prejudice. Indeed, psychologists have long noted that laypeople, like scientists, develop theories as tools to interpret, predict, and control their social world. Fritz Heider (1958), for example, described a “common sense” psychology based on people’s theories about their social world, and in a similar vein George Kelly (1955) called attention to the “personal constructs” that people use to understand their relationships with others and maneuver their way through their social world. These theories have been referred to as lay, implicit, or naive theories because (a) they lack the rigor of scientific theories, and (b) people are not necessarily able to articulate their theories or appreciate the impact of their theories on their social understanding (D.T.Wegener & Petty, 1998; see Hong, Levy, & Chiu, 2001). Yet, despite the fact that they tend to operate in the background, such theories provide crucial constraints on the potentially limitless ways that incoming information might be interpreted and encoded (e.g., G.T. Kelly, 1955; Levy et al., 2001). A growing body of research has focused on identifying particular lay theories and spelling out these theories’ effects on cognition, affect, motivation, and behavior (e.g., Dweck, Chiu, & Hong, 1995; Epstein, 1989; Heider, 1958; Kelly, 1955; Murphy & Medin, 1985; Rhodewalt, 1994; Sternberg, 1985). More recent work has investigated the effects of lay theories on the perception of, and behavior toward, individuals and groups (e.g., Chiu, Hong, Morris, & Menon, 2000; Mendoza-Denton, 1999; Wittenbrink, Hilton, & Gist, 1998; Yzerbyt, Leyens, & Corneille, 1998; see Hong et al., 2001; Karafantis & Levy, 2003). For example, Wittenbrink et al. (1998) demonstrated that different a priori assumptions about the relationships among stereotypic attributes lead perceivers to different sets of inferences about social targets. In short, there is growing recognition of the crucial role played by perceivers’ background theories in the understanding and mental representation of social information.
98
THE PSYCHOLOGY OF GROUP PERCEPTION
In this chapter, we continue in the tradition of this research by contrasting two specific theories that people may hold, either chronically or temporarily, about the nature of human attributes. We provide evidence that, in general, one of these theories (an “entity theory”) is associated with viewing groups as more entitative and more homogeneous, whereas its counterpart (an “incremental theory”) is associated with less rigid expectations of entitativity and homogeneity. In so doing, we outline the assumptions and processing dynamics associated with each lay theory and provide evidence that, in concert with their different ways of representing groups, entity theorists are more likely to endorse stereotypes and overlook information that counters a stereotype, whereas incremental theorists are less likely to endorse stereotypes and more likely to be mindful of counterstereotypic information. Moreover, we develop the notion that these two distinct orientations toward group perception may involve different conceptions of group “essence,” that is, the underlying core nature of groups that binds group members together (e.g., N.Haslam, 1998; Hirschfeld, 1995; Rothbart & Taylor, 1992; Yzerbyt, Rocher, & Schadron, 1997; see also Medin & Ortony, 1989). Finally, we provide evidence that changing these beliefs can directly affect “people’s representations of groups—suggesting a novel and potentially powerful method for reducing stereotyping and prejudice. We now turn to developing each of these ideas in greater detail. Fixed Versus Malleable Theories About Human Attributes In his writings on the epistemology of science, the philosopher Alfred North Whitehead proposed that one metaphysical system, the static worldview, leads naturally to the desire to measure enduring properties and to create taxonomies based on them, whereas another system, the dynamic worldview, leads to the desire to analyze, understand, and influence the processes underlying these dynamic systems (Whitehead, 1929; 1938; see also J.A.Johnson, Gerner, Efran, & Overton, 1988; Pepper, 1942; Unger, Draper, & Pendergrass, 1984). Dweck and colleagues have investigated a related dimension, people’s lay theories about the fixedness versus malleability of human personality, and have found parallel tendencies in person perception (e.g., Dweck, 1999; Dweck, Chiu, & Hong 1995; Dweck & Leggett, 1988). Originating in work on students’ fixed versus malleable conceptions of intelligence, this line of research has identified two primary theories. An entity theory posits that personal characteristics are fixed entities, despite a person’s efforts or motivation to change them. That is, they are not under personal control. (This notion is captured in the statement from one of our scales designed to measure implicit theories, “Everyone is a certain kind of person, and there is not much that can be done to really change that”.) An incremental theory holds that personal characteristics are dynamic and cultivatable with time and effort (as reflected in the statement from one of our scales, “Anyone can change even their most basic qualities.”)1 These two perspectives (entity vs. incremental) have been shown to influence cognition, affect, motivation, and behavior across a wide range of academic, social, and moral domains (for reviews see Dweck, 1999; Dweck, Chiu, & Hong, 1995; Dweck & Leggett, 1988). There is increasing evidence that these two perspectives underlie distinct meaning systems for understanding individuals and groups. How so? An entity theory is about fixed traits and, as such, is strongly associated with expecting a high degree of consistency in people’s behavior over time and across contexts (Chiu et al., 1997; Erdley & Dweck, 1993). When traits are believed to possess considerable meaning and predictive utility, then the task of social perception is to identify and measure people’s fixed traits. In other words, traits become the primary unit of analysis in understanding others (Chiu et al., 1997; Levy & Dweck, 1999; Molden, Plaks, & Dweck, 2001).
IN THE EYE OF THE BEHOLDER
99
In contrast, an incremental theory is about the malleability of personality and behavior. Because of their more dynamic theory, incremental theorists expect less rigid consistency in behavior over time and across contexts (e.g., Chiu et al., 1997; Erdley & Dweck, 1993; Molden et al., 2001). In concert with this more dynamic approach, they are not limited to traits in their understanding of personality and behavior. Indeed, incremental theorists (compared to entity theorists) have been found to give more weight to the mediating psychological processes and situational forces acting on the target (Chiu, 1994; Erdley & Dweck, 1993; Gervey, Chiu, Hong, & Dweck, 1999; Hong, 1994; Molden et al., 2001; Levy & Dweck, 1998, 1999; Levy et al., 1998; Sorich & Dweck, 1996). It is important to note that although the entity and incremental theories can be measured as relatively stable, chronically accessible knowledge structures, they can be altered in various ways. For example, in a number of studies these two theories have been situationally manipulated, yielding similar patterns of findings (suggesting that there is a causal link between the activation of each theory and its associated set of inferences) (e.g., Chiu, Hong, & Dweck, 1997, Study 5; Hong, Chiu, Dweck, Lin, & Wan, 1999, Study 3; Levy & Dweck, 1999; Levy et al., 1998, Study 4; Plaks, Stroessner, Dweck, & Sherman, 2001, Study 3). In one study, Chiu et al. (1997, Study 5) had college students read short versions (either entity or incrementaloriented) of a fictitious Psychology Today type article. Each article cited evidence for the theory from several sources—case studies of individuals (including famous people), longitudinal studies conducted over several decades, and large-scale intervention programs. Some time after reading the article, participants completed a measure that assessed the degree to which they made trait inferences from behaviors. Results indicated that those who read the article supporting an incremental theory made weaker trait judgments and predicted less cross-situational consistency than those who read the article supporting an entity theory. Other work has demonstrated that different cultures (Chiu & Hong, 1999), sociopolitical events (Hong. Chiu, Yeung, & Tong, 1999), and socialization experiences (Mueller & Dweck, 1998) can all similarly alter or inculcate either theory. As we discuss later, the fact that an incremental theory can be activated or fostered in most people is generally encouraging news for researchers interested in reducing stereotyping and prejudice. DO LAY THEORIES INFLUENCE HOW GROUPS ARE PERCEIVED? Do these two models for understanding human behavior, one based on traits and one based on psychological processes, translate into distinct takes on group perception? Do these different lay theories systematically influence perceptions of group entitativity and variability? Research strongly suggests that the answer is yes. We now turn to several converging lines of evidence indicating that entity theorists’ traitbased emphasis is associated with the tendency to (a) apply traits to (and “entitize”) groups, (b) perceive less within-group variability with respect to possessing a trait and greater between-group difference, (c) form stronger stereotypes, (d) selectively avoid counterstereotypic information. Applying Traits to Groups Given that people holding an entity theory invest traits with considerable meaning, Levy, Stroessner, and Dweck (1998) predicted that entity theorists (compared to incremental theorists) would be likely to use traits to characterize an entire group. For example, in one study, Levy et al. (1998, Experiment 3), provided college students with information about a novel (fictitious) group of students ostensibly from another university. The group was characterized by 6 neutral behaviors (e.g., “bought a magazine from a newsstand”) and either 12 positive behaviors (e.g., “ran after a person who left a package”) or 12 negative
100
THE PSYCHOLOGY OF GROUP PERCEPTION
behaviors (e.g., “pushed to the front of the line at the movie theater”). Each behavior was said to have been performed by a different group member. After receiving the information, participants generated an openended description of the group as a whole. Although traits made up much of the open-ended descriptions, entity theorists generated significantly more traits and, additionally, tagged the traits with more extreme qualifiers (e.g., very, extremely) than did incremental theorists. This finding is noteworthy in light of Hamilton and Sherman’s (1996) now-classic reformulation of the role of entitativity in social perception. Hamilton and Sherman (1996) proposed that, contrary to prevailing wisdom, people do not use distinct sets of processes when forming impressions of individuals vs. groups. Instead, people simply expect individual targets to be more internally consistent and coherent than group targets, which, in turn, leads to (a) more extreme and confident dispositional inferences about individual targets, (b) more clearly organized representations of individual targets, (c) more definitive evaluative and causal analyses for individual targets. We suggest that, while people in general may perceive individuals to be more entitative than groups, entity theorists (compared to incremental theorists) tend to perceive higher entitativity in both individuals and groups. After all, when entity theorists apply traits so readily and extremely, this strongly im plies that they perceive the target (whether individual or group) to be a coherent, bounded entity. Indeed, entity theorists may perceive targets to be so entitative that they consider sparse behavioral information to be sufficient for revealing a group’s underlying nature. To examine this hypothesis directly, Levy et al. (1998, Experiment 3), in the study just described, also asked participants, “Do you feel that you were provided with sufficient information to form an impression of the group?” In other words, participants were asked to consider whether the information provided about a subset of group members (6 neutral and 12 positive or negative behaviors performed by different group members) was sufficient for forming an impression of the entire group. Entity theorists, as predicted, felt that the information was appreciably more sufficient than did their incremental theorist counterparts. Other studies have shown that even outside of the context of a judgment task, entity theorists more strongly endorse, and in fact agree with, the following general statements: “Each person has a basic character, and you can tell what kind of person someone is even by details of their behavior or appearance”; “A single act often tells you a lot about a person’s fundamental character”; and “It’s fairly easy to tell what kind of a person someone is by observing them on one or two occasions.” In contrast, incremental theorists’ mean response to these statements fell on the “disagree” side of the rating scale (Sorich & Dweck, 1996). Does this tendency to use traits more frequently and more extremely to characterize groups have any relation to entity theorists’ view of group variability? It is certainly a possibility. Applying a trait term to a group implies that the group is a coherent entity and carries the expectation that group members will behave similarly across situations and over time. In other words, the behavior of different group members is seen as constrained by the trait they presumably all share. This could well translate into perceptions of low group variability. In the next section we describe studies that have investigated this possibility. In contrast, incremental theorists, who tend to rely less on broad group traits and more on dynamic psychological processes and states, should expect (and actually perceive) greater variability among group members.2 Perceptions of Group Variability Levy et al. (1998; Experiment 3) investigated entity and incremental theorists’ perceptions of group variability in the study described above in which participants learned about a novel group. The results demonstrated that, despite a wide range in the group’s behavior, entity theorists, in line with their more extreme trait judgment, judged the group to be significantly more internally similar than did incremental
IN THE EYE OF THE BEHOLDER
101
theorists (see also Levy & Dweck, 1999). Additional studies showed that this entity-incremental difference was driven by entity theorists’ heightened perceptions of within-group similarity rather than incremental theorists’ failure to perceive such similarity (Levy & Dweck, 1999). This basic finding was corroborated by Chow (1996), who asked Hong Kong college students to list stereotypes of Mainland Chinese and, as a measure of perceived outgroup variability, asked them to indicate how many out of 100 Mainland Chinese possessed the attributes they listed. Results indicated that entity theorists saw a greater percentage of Mainland Chinese as possessing both positive and negative stereotypical traits than did incremental theorists. Strictly speaking, percent estimation measures, such as those used in these studies, have been found to be better measures of perceived extremity than perceived variability (Park & Judd, 1990). Park and Judd (1990) have noted, however, that when a groups “mean” is perceived to be extreme, then the group’s variability can also be viewed as small. Furthermore, Park and Judd (1990) highlighted a key distinction between variability conceived of as “dispersion” and variability conceived of as “stereotypicality” (i.e., Does the target possess the stereotypic trait or not?). In light of these insights, the Levy & Dweck (1999) and Chow (1996) data seem to suggest that entity theorists perceive more group homogeneity with respect to possessing the stereotypic trait(s) (i.e. most of the group members do possess the trait). Perceptions of Between-Group Differences Past research has shown that social categorization not only promotes exaggerated perceptions of group homogeneity, but also exaggerated perceptions of between-group differentiation (e.g., Tajfel, 1978a). That is, when forming impressions of multiple groups, people tend to draw relatively clear boundaries between the perceived attributes of one group compared to another (e.g., Ford & Stangor, 1992). Because entity theorists traffic heavily in traits, we expected that they would be more likely to differentiate groups along trait dimensions than would incremental theorists. To test this hypothesis, Levy and Dweck (1999) had children learn about students their age from two fictitious schools. At one school, the students performed some positive behaviors (i.e., 6) and a few neutral behaviors (i.e., 3), while at the other school, the students performed some negative behaviors (i.e., 6) and a few neutral behaviors (i.e., 3). Using a procedure established by Bigler (e.g., Bigler, 1995; Bigler, Jones, & Lobliner, 1997), between-group differentiation scores were calculated as the proportion of one group seen as possessing a trait minus the proportion of the other group seen as possessing the same trait. As predicted, students holding an entity theory saw the novel schools as differing more strongly on the relevant traits (e.g., nice, honest, friendly). Additionally, in this study children were asked to compare the groups on characteristics further removed from the behavioral information provided. They were asked to decide whether none, some, most, or all of the children from the two schools (a) like to do the same kinds of things (games, movies) and (b) have the same concerns (worries, wishes). Entity theorists reported that significantly fewer of the children from the two schools would share likes and concerns than did the incremental theorists. This finding indicates that entity theorists thought of these novel groups as quite different not only on the relevant traits but also in aspects of their psychology that were quite removed from the information provided. They may well have been seeing them as entirely different kinds of people with very little in common. (It is important to note that differences in entity and incremental theorists’ judgments of intragroup similarity and intergroup difference are not due merely to differences in retention of the provided behavioral information [Levy & Dweck, 1999; Experiment 2; Levy et al., 1998, Experiment 3]. Instead, entity and incremental theorists appear to interpret and mentally represent group behavior in distinct ways.)
102
THE PSYCHOLOGY OF GROUP PERCEPTION
In sum, an entity model has been found to promote perceptions of low within-group variability with respect to traits and high between-group difference. In fact, this pattern has recently been replicated with group stereotypes in an Asian culture (in Hong Kong: Hong, Chiu, Yeung, & Tong, 1999). Taken together, these data suggest that perceivers’ construals of group variability are determined to a significant extent by which model (entity or incremental) is active in their minds. Thus, rather than assuming that all perceivers cluster around a single representation, we suggest that different lay theories lead to systematically different representations. Thus far, we have shown that entity theorists are (a) more likely to believe that traits are useful and valid ways of characterizing others and (b) more likely to perceive low variability among group members with respect to having a particular trait or traits. The studies described so far have involved novel groups, about which participants had no prior knowledge. Do these two tendencies translate into a greater tendency to use and believe in existing stereotypes about known groups in the world? It would seem likely, given previous research showing that stereotyping (of known groups) involves (a) attributing a set of fixed traits to an entire group (e.g., Allison & Messick, 1985; Hewstone, 1990; Pettigrew, 1979), and (b) believing in low within-group variability with respect to core traits (e.g., Ford & Stangor, 1992; Sedikides & Ostrom, 1993). In the next section we describe studies that tested this hypothesis. Differences in Stereotype Use and Endorsement In two studies (Levy et al., 1998, Experiments 4 and 5), college students were asked to indicate the extent to which 15 positive and negative traits applied to members of three ethnic groups (African Americans, Asians, and Latinos) and four occupational groups (teachers, doctors, lawyers, and politicians). All groups were rated on all traits, but different traits were stereotypical and nonstereotypical for different groups. The results revealed that entity theorists more strongly endorsed both positive and negative traits relevant to the stereotypes of the ethnic and occupational groups than did incremental theorists. Entity theorists, however, did not differentially endorse traits irrelevant to societal stereotypes of these groups. Significant differences in endorsement of positive and negative stereotypes were also found in a study (Levy et al., 1998, Experiment 1) in which college students were asked to generate a list of stereotypes for five ethnic groups (African Americans, Asians, Caucasians, Hispanics/Latinos, and Jews) (cf. Devine, 1989, Study 1; Eagly, Mladinic, & Otto, 1994). Participants were then instructed to review the stereotypes they listed and to evaluate how true they thought each one was. As predicted, entity theorists endorsed both positive and negative stereotypes of these groups significantly more strongly than did incremental theorists. Additional findings ruled out several potential alternate explanations for our effects. First, the observed differences in agreement with stereotypes could not be attributed to differences in knowledge of societal stereotypes, as entity and incremental theorists generated the same number and type of stereotypes for the provided groups (cf., Devine, 1989; Lepore & Brown, 1997). In addition, they agreed on the evaluative connotations of stereotypes. That is, entity theorists and incremental theorists gave highly similar positivity/ negativity ratings of negative stereotypes (lazy, cheap, and ethnocentric), positive stereotypes (athletic, musical, and hardworking), and neutral stereotypes (food preferences and family characteristics). Moreover, possessing a chronic entity theory has been found to be statistically independent of variables thought to be associated with prejudice, such as right-wing authoritarianism and the need to evaluate (Altemeyer, 1988; Jarvis & Petty, 1996; Levy et al., 1998, Experiment 5). This is not surprising considering that an entity theory is simply a belief about the fixedness or malleability of human personality and contains nothing that directly implies more extreme views about particular groups or more extreme views in general.
IN THE EYE OF THE BEHOLDER
103
These data indicate that the differences in stereotyping exhibited by entity and incremental theorists are not due to differences in knowledge of stereotypes, motivation to understand the target, a general tendency to respond more extremely, or prejudice. Instead, each theory seems to precipitate a different framework for mentally representing groups and understanding their behavior. Moreover, there is reason to believe that not only do entity and incremental theorists differ in their representations of individuals and groups, but they also adopt distinct modes of processing that facilitate and bolster their respective representations. We turn to this topic next. Selective Attention Underlying Each Perspective Thus far, we have described how the entity framework, compared to the incremental framework, appears to lead to a greater tendency to characterize groups with traits, high consistency with respect to possessing those traits, and a greater proclivity for stereotyping existing groups. Are there differences in underlying cognitive processes that might bolster and maintain such differences in inference and judgment? We suggest that if these theories truly underlie distinct processing frameworks, then entity and incremental theorists might differ not only in their inferences and judgments about targets, but also in how they attend to and encode target information (e.g., Hong, Chiu, Dweck, & Sacks, 1997; McConnell, 2001; Plaks et al., 2001). Moreover, such “earlier” processes (attention, encoding) may facilitate reaching judgments that are consistent with their theory. In this section, we describe a series of studies concerning such mediating, attentional processes. How do people respond when faced with a target who acts in a clearly counterstereotypic manner (e.g., a female construction worker)? Do entity and incremental theorists differ in their responses? Answers to these questions would have noteworthy implications for perceived group variability and the maintenance of stereotypes. Someone who discounts or ignores such discrepant information may preserve a representation of low group variability, thereby preserving the integrity of the stereotype. In contrast, someone who mentally acknowledges and incorporates such variability within a group may come to revise or even discard the group stereotype due to its apparent lack of predictive power. The prior literature on this question is ambiguous. Some studies have suggested that people allocate more attention to consistent information due to its comparative ease of processing (e.g., Bodenhausen & Lichtenstein, 1987; Fiske & Neuberg, 1990; Hilton, Klein, & von Hippel, 1991). This would have the natural consequence of helping to perpetuate representations of high group homogeneity. On the other hand, other studies indicate that perceivers process inconsistent information with greater elaboration than consistent information (in attempting to understand or reconcile the inconsistency), as in the well-known “incongruency effect” (e.g., Hastie & Kumar, 1979; Sherman, Lee, Bessenoff, et al., 1998; Srull, 1981). Greater attention to counterstereotypic information can, in theory, weaken representations of group homogeneity. Past research has also begun to investigate conditions under which people prefer one route over the other (e.g., Erber & Fiske, 1984; Stangor & Ruble, 1989). Recent research in our laboratory addresses this issue. Research on Lay Theories and Selective Attention Based on the growing evidence that entity theorists use and endorse stereotypes to a greater extent than do incremental theorists, we (Plaks et al., 2001) hypothesized that entity theorists would treat counterstereotypic information in a manner that encourages stereotype preservation. One possible way to accomplish this is to minimize exposure to inconsistent information relative to consistent information.
104
THE PSYCHOLOGY OF GROUP PERCEPTION
By regulating the entry of evidence of group members behaving counterstereotypically, an impression of high group homogeneity can be maintained. In contrast, incremental theorists were hypothesized to exhibit the typical incongruency effect found in the literature, presumably because their dynamic view of personality would lead to interest in inconsistent information’s novel, challenging, or diagnostic qualities (e.g., Bassok & Trope, 1984; Hastie & Kumar, 1979; Macrae, Bodenhausen, Schloerscheidt, & Milne, 1999; Sherman, Lee, Bessenhoff, et al., 1998; Stangor & McMillan, 1992). In one study, an attentional probe paradigm was used to measure participants’ on-line attention to stereotype-consistent and -inconsistent information (Plaks et al., 2001, Experiment 1). With this technique it is possible to measure on-line attention to each type of information without participants’ awareness (e.g., Hashtroudi, Mutter, Cole, & Green, 1984; Sherman et al., 1998). Participants read a series of sentences, one at a time, describing the behavior of someone labeled as either a priest or a neo-Nazi skinhead. One-third of the sentences featured kind behaviors (e.g., “Robert gave up a taxi to a stranger”), one-third featured unkind behaviors (e.g., “Robert pushed to the front of the line at the movie theater”), and one-third featured behaviors that were irrelevant to the kind/unkind dimension (e.g., “Robert took the bus downtown”). (Naturally, the kind behaviors were considered to be consistent with the priest stereotype and inconsistent with the skinhead stereotype, and the unkind behaviors were considered to be consistent with the skinhead stereotype and inconsistent with the priest stereotype.) During the presentation of nine of the sentences (three from each of the three types of sentences: kind, unkind, and irrelevant), the computer emitted a tone, and participants were instructed to press the space bar as quickly as possible after hearing the tone. Prior studies have demonstrated that when people are highly engaged with the stimulus that is currently in the visual field, their reaction time to a concurrent tone is retarded, compared with people who are not as engaged (e.g., Hashtroudi et al., 1984; Sherman, Lee, Bessenoff, et al., 1998). Participants’ lay theory was measured in a later session during what was ostensibly part of another study. As predicted, entity theorists exhibited slower reaction times (indicating less attention) to inconsistent information than to consistent information or irrelevant information, whereas incremental theorists showed the opposite pattern (Plaks et al., in press, Experiment 1).3 Thus it appears that lay theories systematically direct attention allocation in distinct manners that are consistent with the mental models implied by each theory. Given findings from the entitativity literature that people expect individuals to behave more consistently than groups (e.g., Susskind, Maurer, Thalekar, Hamilton, & Sherman, 1999), would the pattern we found with individual targets transfer to group targets? Would entity theorists avoid evidence of group inconsistency or variability? Would incremental theorists approach such information? To test this, in a second study (Plaks et al., 2001; Experiment 3), we specifically employed group targets. Rather than reading that “Robert” performed each kind, unkind, or irrelevant behavior, participants were told that a different member of the group performed each behavior. The results replicated the pattern found in Experiment 1. Participants with an entity theory (this time, temporarily primed, see Chiu et al., 1997, Experiment 5; Levy et al., 1998, Experiment 4) allocated less attention to counterstereotypic information than to stereotypic information and irrelevant information. Participants exposed to the incremental theory induction displayed the reverse pattern. Thus, regardless of whether the target is an individual or an entire group, and regardless of whether theories are measured or manipulated, entity theorists’ attention allocation seems aimed at preserving a consistent, homogeneous view of the target, whereas incremental theorists’ attention allocation seems more open to inconsistency and variability. In fact, recent data provide more direct evidence of the proposed connection between participants’ attention to target information and the impressions they form. Using a dichotic listening task, Plaks et al.
IN THE EYE OF THE BEHOLDER
105
(2001, Experiment 4) found that entity theorists diverted their attention away from a target as his behavior became increasingly counterstereotypic. (The attention of incremental theorists, in contrast, was not influenced by the stereotypicality of the target’s behavior.) In Part 2 of the study, participants were asked to render judgments about this counterstereotypically-acting target. Plaks et al. (2001) found a notable divergence: Incremental theorists’ impression of the target changed in a linear fashion, in direct proportion to the amount of counterstereotypic behaviors performed by the target. Entity theorists, on the other hand, exhibited a strikingly different pattern: In the face of small to moderate amounts of counterstereotypic behavior, their judgments of the target remained constant. This rigidity of the impression, combined with their decrease in attention, was interpreted by Plaks et al. (2001) as evidence that entity theorists “screened out” a portion of the counterstereotypic information, successfully preventing this information from impacting their judgments. Only in a condition in which the target’s behavior was overwhelmingly counterstereotypic did entity theorists’ impressions of the target finally undergo a significant revision. Thus entity theorists appeared to use selective attention to cling to their stereotypic impression of the target for as long it was realistically feasible. This link between information intake and judgment is further supported by recent work by McConnell (2001) showing that entity theorists, when forming an impression of someone, use a more “on-line” mode of processing whereas incremental theorists use a more “memory-based” mode (Hastie & Park, 1986). These different modes of processing lead to strikingly different judgments based on the same presented information. According to McConnell’s account, entity theorists’ on-line processing reflects a continual effort to harmonize the disparate pieces of target information into a single trait construct (cf. Hong, Chiu, Dweck, & Sacks, 1997). Incremental theorists are presumably not as concerned with categorizing the target in trait terms, and therefore perform less on-line elaboration of incoming information. Why exactly does the entity framework favor trait- and stereotype-consistent information over inconsistent information? Is it simply because entity theorists consider consistent information more diagnostic of the target’s “true” personality (and inconsistent information mere “noise”)? Or might it also be that the inconsistent information represents a threat that activates defensive motivations? Recent evidence indicates that defensive motivations may play an important role. First, in several studies, entity theorists not only paid less attention to inconsistent than consistent information, but they also paid less attention to inconsistent than irrelevant information (e.g., Plaks et al., 2001, Experiment 1, Experiment 3). If diagnosticity were their only concern, then inconsistent and irrelevant information, by both being “not consistent,” should attract equal attention. The finding that entity theorists repeatedly paid less attention to inconsistent than irrelevant information may reflect a motivation to avoid inconsistent information, relative to the neutral baseline. Moreover, in another study, in which the amount of counterstereotypic behavior was varied, entity theorists (but not incremental theorists) steadily decreased their attention to a target as the target’s behavior became increasingly counterstereotypic (Plaks et al., in press, Experiment 4). Finally, in three recent studies, the more attention entity theorists paid to certain counterstereotypic information, the greater the rise in self-reported anxiety (Plaks, 2001, Experiment 2; Plaks, Grant-Pillow, & Dweck, 2002). This relationship was not found for incremental theorists. Taken together, these studies provide support for a significant defense-motivational component to entity theorists’ preference for trait- and stereotypeconfirming information.4 In summary, these and other data (Plaks, Grant-Pillow, Dweck, 2002) suggest that entity theorists’ general tendency to seize upon stereotype-confirming information at the expense of counterstereotypic information may contribute to the maintenance and even strengthening of stereotypes, perhaps providing even stronger evidence for the group’s homogeneity. In contrast, the attention allocation of incremental theorists appears more open to inconsistent information (which could, in principle, lead more easily to
106
THE PSYCHOLOGY OF GROUP PERCEPTION
stereotype revision or to impressions of higher within-group variability). Though incremental theorists clearly do invoke stereotypes, they may view stereotypes as provisional labels that are open to revision. The findings reported in this section extend previous findings demonstrating that entity theorists readily apply traits to groups (e.g., Levy et al., 1998), and bolster our contention that the entity-incremental variable can complement inherent target variables (e.g., entitativity) in determining people s ultimate impression of a group. FURTHER IMPLICATIONS FOR PERCEPTIONS OF GROUP HOMOGENEITY, ENTITATIVITY, AND ESSENCE Thus far, we have provided evidence that, compared to people operating with an incremental model, those with entity model tend to (a) represent others as entitative units comprised of fixed traits, (b) perceive low within-group variability, (c) register more agreement with stereotypes, and (d) engage in selective attention that bolsters perceptions of group homogeneity and encourages stereotype confirmation. With respect to item (a), we have argued that entity theorists’ proclivity for describing groups using trait labels reflects a tendency to view groups as coherent, entitative units. A basic assumption of “groupness” would appear to be a precondition for invoking strong traits and harboring expectations of high group homogeneity. In addition, at least two pieces of evidence further support the notion that entity theorists are prone to viewing groups as discrete, coherent entities. First, entity theorists tend to demarcate the boundaries between groups more clearly than do incremental theorists (Levy & Dweck, 1999). Second, entity theorists tend to extend traits even to unknown, unencountered group members (Levy & Dweck, 1999), seemingly reflecting an acute sense that the group in question is a phenomenologically “real” entity. Lay Theories and Perceptions of Essence Is there a single concept that can tie together this pattern of findings? We view the concept of essence as a good candidate. Essence refers to an underlying core nature of a group, which binds group members together (e.g., N.Haslam, 1998a; Hirschfeld, 1995; Rothbart & Taylor, 1992; Yzerbyt, Rocher & Schadron, 1997; see also Medin & Ortony, 1989). According to Haslam and colleagues (Haslam et al., this volume), essentialist beliefs reflect assumptions about two factors: (a) entitativity (or “reification”) and (b) shared biology among social group members (see also Hirschfeld, 1996; Rothbart & Taylor, 1992). Endorsing such beliefs has been shown to be associated with more extreme group dispositional judgments and other markers of stereotypic thought (Rogier & Yzerbyt, 1999; Yzerbyt et al., 1998; Yzerbyt, Rocher, & Schadron, 1997). We have presented evidence that entity theorists tend to reify groups, based on traits, to a greater extent than do incremental theorists. What about the second, genetic component of Haslam et al.’s (this volume) formulation of essence? Recent evidence indicates that, not only do entity theorists endorse traits more strongly than do incremental theorists, but they tend to attribute the presence of group traits more to innate factors and less to shared environment and experiences. In one study (Levy et al., 1998, Experiment 2), participants were asked to rate the extent to which stereotypic views of African Americans existed or persisted because of “innate factors” and “past or present environmental or social causes within American society,” as well as three filler explanations. The results indicated that entity theorists believed to a greater degree than incremental theorists that these stereotypic traits are innate. Incremental theorists, relative to entity theorists, saw the causes of group stereotype persistence more in terms of social and environmental factors.
IN THE EYE OF THE BEHOLDER
107
This greater belief in the innate aspect of traits means that entity theorists are more likely to endorse both aspects of essence. Indeed, coherence around key biologically based traits may be, according to the entity view, what defines many groups. In line with these assumptions (and consistent with past work on essence), more extreme trait ratings and more extreme perceptions of homogeneity may follow suit. Different Mental Models of Human Behavior It is important to note that the entity and incremental theories are not theories about group entitativity, homogeneity, or essence per se. Instead, they are theories regarding the general characteristics of human attributes and behavior. Yet each theory readily implies a definable set of inferences about groups. Thus, it appears that the activation of each theory in turn activates a network of inferences and allied beliefs that are likely to be “true” if the core belief is “true.” In other words, each theory may lay the foundation for a distinct mental model of human behavior and of what makes a group of people a “group.” (For a more detailed treatment of this mental models approach to lay theories, see Levy, Plaks, & Dweck, 1999). As noted, according to the entity model, groupness is often defined with respect to biologically shared traits. What about the incremental model? How do incremental theorists define groupness if not through traits? One intriguing possibility implied by our findings is that the incremental model sees groupness as defined by shared dynamic, psychological constructs such as goals, attitudes, or experiences. Prior research has alluded to both traits and psychological processes as bases for perceived entitativity (Lickel et al., 2000), although no clear explanation has been proffered for when one dimension might be preferred over another. We suggest that people with a trait-orientation are especially sensitive to entitativity along the shared traits dimension and people with a process-orientation are especially sensitive to entitativity along the shared goals/ experiences dimension. In fact, this possibility is directly supported by Chiu (1994), who found that entity theorists judge the similarity of people based primarily on their traits, whereas incremental theorists judge people as similar based primarily on their shared goals. Moreover, recent data show that incremental theorists are more likely than entity theorists to detect a goal-directed pattern in a group’s behavior across multiple situations (Plaks, Shafer, & Shoda, 2003). Thus, incremental theorists may perceive coherence in groups just as readily as entity theorists do, but based on different criteria. However, because there are, for example, multiple ways to pursue the same goal, there needn’t be as much of an assumption of homogeneity of behavior. In other words, although incremental theorists may entitize groups in the sense of seeing them as coherent, their style of perceived entitativity may not foster the same degree of perceived homogeneity, a possibility that further underscores the distinction between perceived entitativity and homogeneity (Hamilton & Sherman, this volume). These hypotheses await future research. In sum, groups clearly differ in the extent to which they naturally afford perceptions of homogeneity, entitativity, and essence (Lickel, et al., 2000). Yet we have proposed that, rather than there being one objective “score” for every group, perceptions of entitativity, homogeneity, and essence are highly dependent on the different assumptions and theories about human nature that are most accessible in the perceiver’s mind. Thus, an understanding of differences in perceivers’ a priori beliefs can complement knowledge of the inherent qualities of groups. By including both ends of the equation, one can account for a greater proportion of the variability in group perception and make sharper predictions about perceivers’ inferences and behavior toward group members.
108
THE PSYCHOLOGY OF GROUP PERCEPTION
Implications for Prejudice Reduction and Intergroup Relations As noted, the entity and incremental models can be either chronically accessible or situationally induced (e.g., Chiu et al., 1997, Experiment 5; Levy et al., 1998, Experiment 4; Plaks et al., 2001, Experiment 3). This finding is important because it suggests that people understand both theories. Therefore, teaching or heightening the accessibility of an incremental framework may be an accomplishable and effective means of reducing stereotypic thought and its symptoms. Indeed, recent research bears this out; in several studies, adult and child participants who were exposed to a situational incremental theory induction indicated less agreement with stereotypes (Levy et al., 1998, Experiment 5), made less extreme trait judgments regarding groups, expected greater within-group trait variability, did not extend trait judgments to unknown group members (Levy, 1998), and paid more attention to countersteotypic information (Plaks et al., 2001, Experiment 3). These incremental theory inductions were brief, onetime affairs; presumably a sustained educational program would achieve even more effective and long-lasting effects. These findings suggest that rather than attempting to debunk people’s stereotypes about particular groups one by one, a more effective and meaningful change in stereotypic thinking may involve activating an incremental theory. If intergroup interactions are to succeed in debunking stereotypes, participants must be prepared to acknowledge counterstereotypic information and evidence of within-group variability when they see it. It appears that an incremental theory, through its emphasis on dynamic psychological processes, may foster this, by orienting people toward acknowledgment and appreciation of the variety of human behavior. In fact, recent findings suggest that people holding an incremental view actually seek out experiences that involve bridging diverse social groups and finding intergroup commonalities. In several studies that examined students’ volunteering behavior, students with an incremental view had volunteered more in the past, were more active volunteers, and were more willing to volunteer in the future to help the disadvantaged groups (presumed outgroups; Karafantis & Levy, 2003; West & Levy, 2003). Greater experience interacting with outgroup members would provide opportunities to learn about within-group variability and between-group commonalities. When Might Incremental Theorists Show More Prejudice? We have argued here and elsewhere (e.g., Levy et al., 1998; Plaks et al., 2001) that entity theorists are not simply more prejudiced than incremental theorists. Indeed, the incremental framework might sometimes lead to prejudice and outgroup derogation, albeit in a different way. How so? The belief that human attributes are cultivatable over time implies that, with effort, people have control over what they become. This belief could lead incremental theorists to view a group negatively (a) when members of the group are perceived as choosing to become something considered to be socially undesirable or (b) when there is the perceived risk that group members might socialize others into a socially undesirable outlook or lifestyle.5 These two assumptions may, for instance, lead incremental theorists to exhibit homophobia. That is, some incremental theorists may believe that homosexuals have consciously chosen an undesirable lifestyle and may fear that homosexuals can attract others into developing that side of their nature and adopting that lifestyle. In fact, Haslam, Rothschild, and Ernst (2002) have found that participants with an incremental theory exhibited significant antigay attitudes.6 At the same time, it is important to note that our perspective is consistent with Begue and Apostolides’ (in press) finding that entity theorists view homosexuals as more different from themselves. Entity theorists, with their trait approach, may place homosexuals in a different category than themselves or other heterosexuals. This mental segregation of the outgroup is in accord with entity theorists’ greater stereotyping and labeling, but may also lead to less fear that the outgroup’s attributes will be adopted by others.
IN THE EYE OF THE BEHOLDER
109
In sum, although we have focused in this chapter on the ways in which holding an entity theory leads people to more extreme use of traits and stereotypes, our broader perspective deals with the ways in which the meaning system people adopt leads them to conceptualize and judge groups in a characteristic manner. CONCLUSION In this chapter, we have proposed that rather than there being one “gold standard” representation for every group around which people cluster, such representations are in the eye of the beholder: Differences in group representation covary meaningfully with the perceiver’s lay theory. Lay theories consistently influence the extent to which people see a given group as a coherent, homogeneous entity whose members share traits. In short, lay theories appear to play a pivotal role in the mental representations of groups, in the processing of group-related information, and thus in the formation and maintenance of stereotypes. It is clear, then, that factors within the perceiver merit much continued research attention given the theoretical and practical importance of understanding the roots of stereotyping. NOTES 1. Participants’ responses to these types of items are used to gauge their most chronically accessible theory. Categorization of participants as chronic entity or incremental theorists is based on an index of agreement or disagreement with the items. Across dozens of studies, approximately 80–85% of the respondents are categorized clearly (and fairly equally) into entity and incremental theorists while the remaining 15– 20%, who show no clear pattern of agreement or disagreement, are typically excluded. In addition, in several studies we have situationally activated each theory (e.g., Chiu, et al., 1997, Experiment 5; Levy et al., 1998, Experiment 4; Plaks, Stroessner, Dweck, & Sherman, 2001, Experiment 3), yielding similar findings as when theories were measured. 2. At least, as we describe below, when the putative similarity is based on shared traits. 3. As in prior research (Sherman, Lee, Bessenoff, et al., 1998), this effect was found only under conditions of high cognitive load. This is consistent with Sherman et al.’s model of encoding flexibility, which states that when people are under high cognitive load, stereotypes work by liberating resources that people may then allocate toward information considered most helpful. For incremental theorists, it appears that inconsistent information fits this bill, but for entity theorists, consistent information may seem more meaningful or appealing. 4. Stereotype preservation, like many basic features of our mental system, is likely to be multiply determined. The Plaks et al. (2001) studies demonstrate that one way to accomplish stereotype preservation is to screen out counterstereotypic information. Another method is to confront and debunk such information. Most likely, the first method will occur when (a) people are under multiple task demands, (b) the counterstereotypic information does not possess a basis for debunking. The second method is more likely when (a) people can concentrate on the task of debunking and (b) the material is rich enough to provide a basis for debunking. Indeed, a recent study (Plaks, Grant-Pillow, & Dweck, 2002) has demonstrated that both entity and incremental theorists may, under the appropriate circumstances, address the presence of motivationally arousing, target information through increased (rather than decreased) attention (see Forster, Higgins, & Strack, 2000; Yzerbyt et al., 2000 for similar findings). 5. It should also be noted that McConnell (2001) has demonstrated that, when forming an impression, an incremental theory is associated with a greater incidence of illusory correlation. This tendency, however, should be distinguished from prejudice. 6. It should be noted, however, that Haslam et al. measured implicit theories with respect to category immutability rather than attribute immutability, as we have done in our laboratory. Indeed, Levy et al. (2001) found that incremental theorists display less prejudice toward marginalized groups with permeable boundaries, (e.g., the homeless) when implicit theories are measured with respect to attributes rather than categories.
8 Components of Homogeneity A Multiple-Process Model of Social Categorization KARL CHRISTOPH KLAUER KATJA EHRENBERG INGO WEGENER Rheinische Friedrich-Wilhelms—Universität Bonn
According to Tajfel (1969, 1982), the act of social categorization leads to the accentuation of intergroup differences and intragroup similarities, thereby increasing the exchangeability of members of the same category and decreasing the similarities between members of different categories. Research on this accentuation principle soon revealed an interesting asymmetry in favour of outgroup homogeneity. That is, people judge members of outgroups as more similar to one another than they do members of ingroups. Perceptions of group homogeneity have implications for individual group members. Members of groups that are perceived as homogeneous may be viewed as exemplifying the group’s central tendency, especially in the absence of individuating information. People may thereby feel justified in letting themselves be guided by stereotypical beliefs about the group in question in judging and behaving toward individual group members. Most studies of group homogeneity have used one of three kinds of measures (Park & Judd, 1990). These are stereotype endorsement, perceived dispersion, and perceived similarity. Measures of stereotype endorsement relate to perceptions of the extent to which group members conform to the group stereotype. Measures of perceived dispersion focus on the perceived variability of group members on specific traits. Finally, perceived similarity is normally measured by having respondents rate the overall similarity of group members. While these variables rely on explicit judgments, a number of studies attempted to assess homogeneity by the confusability of members of the same category in memory tasks such as S.E.Taylor, Fiske, Etcoff, and Ruderman’s (1978) so-called “Who said what?” paradigm (e.g., Judd & Park, 1988; Lorenzo-Cioldi, 1993; Lorenzo-Cioldi, Eagly, & Stewart, 1995; Ostrom, Carpenter, Sedikides, & Li, 1993). In the “Who said what?” paradigm, participants perceive a discussion between members of a small group. The discussants are members of two or more social categories. For example, there may be men and women. Having observed the discussion, participants are shown the individual discussion statements and are asked, in a surprise
COMPONENTS OF HOMOGENEITY
111
memory test, to assign each statement to its speaker. Of prime interest are the assignment errors. It is often found that assignment errors tend to respect the category boundaries. For example, a man’s statement is more often assigned to another man than to a woman when an erroneous assignment is made, even after a suitable chance correction for guessing (cf. S.E.Taylor et al., 1978). The difference of within-category errors and between-categories errors is thus positive and indicates that social categorization spontaneously occurs and that social-category memberships are encoded in memory and used in subsequent retrieval attempts. The within-category errors in particular rather directly assess the confusability of members of the same category, and the error-difference measure relates that confusability to that of members of different categories. Homogeneity is given to the extent to which there are many within-category confusions, but to qualify as group homogeneity, confusions must also respect category boundaries. In particular, the withincategory confusions should exceed between-categories confusions. Based on what was originally a methodological criticism, Klauer and Wegener (1998) have proposed to extend the paradigm slightly. In the extended version, new statements or distracters are also presented in the assignment test, and participants first decide whether a given statement already occurred in the discussion or whether it is new. In the former case they are then asked to assign the statement to its speaker as usual. Through the use of a mathematical model, Klauer and Wegener showed that six distinct cognitive processes determine the assignment patterns in the assignment task. The mathematical model is a member of the class of multinomial models that have seen a growing range of successful applications in recent years (Batchelder & Riefer, 1999). The six identified processes are: 1. Item memory: Does the participant remember the statement that he or she is about to assign? Strength of item memory is measured by the model parameters DA, DB, according to the speaker’s social category (A versus B), and DN, where N denotes new items. DN refers to the likelihood with which distracters are detected. 2. Person memory: Given item memory, does the participant recall the speaker of the statement? Strength of person memory is measured by the model parameters cA and cB for speakers from Category A and B, respectively. 3. Category memory: Given item memory, but no person memory for a statement, is the speakers’s category membership still available and used in assigning the statement? Strength of category memory is measured by the model parameters dA and dB for speakers from Category A and B, respectively. 4. Guessing the category membership: If the information retrieved for a given statement is not sufficient for determining the speakers’s category membership, category membership is guessed, based on whatever information is available at retrieval and the statement is then assigned to a member of that category. The strength of the tendency to guess Category A rather than B is measured by the model parameter a. a=.5 means that A and B are guessed equally frequently, whereas a>.5 means that there is a bias in favor of Category A, and a<.5 a bias in favor of Category B. 5. Guessing the item status: If the information retrieved is not sufficient for determining whether the item was presented at all, the item status is guessed. The strength of the tendency to guess old rather than new is measured by the model parameter b. 6. Guessing the speaker within the correct category: If the speaker’s category, but not his or her identity, can be recollected or reconstructed, the correct speaker is guessed with probability of success given by 1/n, where n is the size of the category.
112
THE PSYCHOLOGY OF GROUP PERCEPTION
According to this analysis, group homogeneity as expressed in the “Who said what?” paradigm is driven by different process components. Let us briefly consider the contributions of person memory, category memory, and category guessing. When person memory is high, the remembered statements are strongly associated with their individual speakers. The statements thereby constitute distinct features of speakers that enhance their discriminability. Assigning a statement to the correct speaker due to person memory requires discriminability of the persons as a necessary precondition. To the extent to which there is person memory, the likelihood of erroneous assignments will be reduced, hence the amount of confusions as well as the error-difference measure will tend to be reduced in size. On the other hand, an antagonistic tendency is induced by category memory. When category memory is high, the remembered statements are strongly associated with all speakers of a given category. The statements thereby constitute common features of these persons enhancing their similarity. As a consequence, the likelihood of within-category confusions will be increased, and the likelihood of betweencategory confusions reduced. Category memory contributes a direct and strong effect on the within-category confusions and the error-difference measure. Finally, when there is neither person memory nor category memory, reconstructive category guessing can indicate perceived group homogeneity of a subtly different kind. Assume that statements were presented that were either pro or contra a given issue (e.g., abortion). When uncertain about a statement’s origin, participants may use statement content to guess a likely category source provided that they believe that attitude on the issue discussed and category membership covary. Perceivers are then likely to guess that the statement stemmed from the category with the attitude that is more consistent with it. Expectancies linking attitudes with categories can be based on stereotypes or can be generated online as the discussion unfolds. In either case, they will lead to more within-category confusions and a larger error-difference measure to the extent to which the expectancy conforms to the realities of the discussion. However, even when there is in fact no covariation between statement content and category membership in the presented discussion, the category guessing parameters nevertheless allow one to assess whether group-level expectancies drive participants’ assignments. It is interesting to relate these different process components of group homogeneity to traditional judgment-based measures of group homogeneity and to theoretical accounts of group homogeneity. For example, the extent of person memory is clearly the causal factor underlying Wilder’s (1984, 1986) account in terms of degree of intimacy of interaction with group members. According to that account, perceptions of group homogeneity will be reduced when the quality of interaction allows us to know and cognitively store the unique and individuating characteristics of each group member. Alternatively, the extent of category memory is most closely linked to measures of perceived similarity according to the above discussion, and it has ties to accounts that assume differential memory representations for groups that differ in homogeneity; homogeneous groups being characterized by more abstract category-level representations (e.g., Mullen, 1991; Ostrom & Sedikides, 1992; Ostrom et al., 1993). Finally, as elaborated below, the parameters assessing category guessing provide measures of stereotype endorsement when the discussion topic is stereotype-relevant. In any case, the extent of reconstructive category guessing is likely to reflect stored beliefs about the extent of group homogeneity (Park & Judd, 1990; Ostrom & Sedikides, 1992; Wilder, 1984) with respect to the category members’ stances on the content issues addressed by the statements. Because of these different theoretical bases of the processes that contribute to observed confusability in the “Who said what?” paradigm, it is likely that the different components of group homogeneity can be affected differently by experimental manipulations. In some cases, they will converge to produce high overall group homogeneity; in other cases, they will exert antagonistic effects. It is the purpose of the
COMPONENTS OF HOMOGENEITY
113
present chapter to explore and analyze this interplay in different experimental contexts. In particular, the effects of category accessibility and fit, of relative group size, and of crossing categories on different components of homogeneity will be examined. In each case, the model analyses permit powerful tests of competing theories about the roles of these variables in person and group perception. The chapter concludes with a discussion of the wider implications of these findings for theories of person and group perception. VALIDATING THE PROCESS MEASURES As a first step, it is important to validate the proposed process measures. In a first series of experiments, we addressed the question of whether the different process parameters are valid measures of their respective cognitive processes. For this purpose, experimental comparisons were implemented according to the following principle: For each of the above-mentioned processes, an experimental manipulation was devised which was known or could be assumed to affect only that process while leaving the other processes unaffected. Then, the model was fit to the data from the different experimental conditions, and it was tested whether differences between conditions were found only in the model parameters measuring the manipulated process but not in parameters associated with other processes. For example, to validate the parameters for category memory, two experimental conditions were compared in Klauer and Wegener’s (1998) Experiment 3 in a between-participants design. Participants in both conditions saw a group discussion of eight persons who made statements in turn. Each speaker made a total of six statements. The discussion revolved around conditions of studying at the university. In one experimental condition, four speakers were described as students from the University of Münster (Category A); the remaining four as students from the University of Aachen (Category B). In the second experimental condition, four speakers were described as students (Category A), and the remaining four as teachers (Category B). It was argued that both experimental conditions would differ primarily in terms of category salience: The categorization on the basis of hometown is likely to be less chronically accessible than that based on academic status in our University of Bonn student population of participants. In addition, the discussion topic, conditions at the university, was relevant with respect to the categorization along academic status, which should further prime the use of this particular categorization and thereby raise its accessibility (Oakes, 1987; Oakes, Haslam, & Turner, 1994). The model fitted the data very well. Figure 8.1 shows the model parameters estimated separately for the two experimental conditions. As can be seen, the categorization according to academic status was highly salient in that it led to high category memory as measured by the d-parameters, while the category memory parameters for the categorization based on hometown were much smaller. In addition, the other kinds of parameters were not significantly affected by the experimental condition that focused on manipulating category salience. This finding established both convergent validity for the proposed measure of category memory as well as discriminant validity for the other kinds of parameters (Campbell & Fiske, 1959). In Experiment 4 of that series, we focused on reconstructive strategies of category guessing during retrieval as measured by parameter a. Four men discussed with four women. In the first experimental condition, men made proportionally more negative statements about conditions at the university than women; in the second experimental condition, women were more critical. Thus, in both conditions there was what has been called comparative fit (Oakes et al., 1994) or cue diagnosticity (Rosch, Mervis, Gray, Johnson, & Boyes-Braem, 1976), that is, a covariation of category membership and kind of statement, but its sign differed between the two groups. The model was fitted with separate parameters for both experimental conditions and kinds of statements (negative versus positive statements). The results for the measures of reconstructive guessing (a-parameters) are shown in Figure 8.2. As can be seen, perceivers
114
THE PSYCHOLOGY OF GROUP PERCEPTION
FIGURE 8.1. Parameter estimates of the multinomial model of social categorization in the modified “Who said what?” paradigm
FIGURE 8.2. Estimates of the guessing parameter: Probabilities of guessing that a statement was made by a woman rather than a man
noticed the comparative fit and used it in filling memory gaps when uncertain: The probability a of assigning a negative statement to a female rather than male speaker was much higher in the condition with critical women than in that with critical men, whereas the reverse was true for positive statements. Again, the other kinds of parameters were not a function of the experimental condition. This pattern of findings suggests that participants noticed the implemented fit and used it in reconstructive category guessing when they were uncertain about the speaker s identity. Similarly, Experiment 5 manipulated the perceptual similarity of the category members. Again, this reduced only the person memory parameter, whereas the other model parameters were not significantly affected.
COMPONENTS OF HOMOGENEITY
115
ACCESSIBILITY AND FIT As already discussed in the introduction, installing comparative fit in the discussion is likely to enhance perceived category homogeneity through group-level expectancies about the homogeneity of the group members’ attitudes. The impact of fit on this component of group homogeneity can be assessed by the parameters for reconstructive guessing as exemplified in the previous section. An interesting question is whether comparative fit also exerts an influence on category memory, that is, whether it also increases the accentuation effect of social categorization in memory. In fact, Oakes (1987) proposed an influential analysis of category salience in terms of accessibility and comparative fit. Building on Bruner’s (1957) work, she argued that category salience is a product of accessibility and fit. Fit as defined by Oakes (1987) is given for a possible categorization to the extent to which there is cue validity (Rosch et al., 1976) for that categorization in the current social context. Like Rosch et al., Oakes believes that the concrete world will tend to be categorized by means of categories for which cue validity is high. Unlike Rosch et al., Oakes holds that cue validity is assessed anew for each actual social context rather than abstracted from many prior experiences and stored in long-term memory. This leads to highly context-dependent social categorizations. In empirical studies that evaluate this accessibility witnesses×fit approach (Blanz, 1999; Oakes, Turner, & Haslam, 1991; van Knippenberg, van Twuyver, & Pepels, 1994), fit has frequently been manipulated as a covariation between the contents of statements made by target persons and the targets’ category memberships. Typically, statements were either pro or con with respect to a discussion topic, and fit was given when members of one category made more pro- than con-statements and vice versa for members of the other category. When accessibility and fit were manipulated in the “Who said what?” paradigm (Blanz, 1999; van Knippenberg et al., 1994), an increase in the traditional error-difference measure of category salience was found as a function of fit. In an unpublished dissertation, Wegener (2000) used the improved paradigm to disentangle the two category-based processing modes—category memory and reconstructive guessing—that may have caused this increase. Wegener (2000) used three different categorizations that were pretested to differ in accessibility: In order of increasing accessibility, they were based on hometown (Aachen versus Münster), professional status (unemployed versus young entrepreneurs), and sexual preference (homosexual versus heterosexual). The category members discussed gender roles in romantic relationships, and discussion statements were of two kinds: conservative or progressive. Orthogonally, fit was manipulated in three steps—no fit, medium fits, and high fit—by increasing the covariation between category membership and kind of statement. In the condition of no fit, there was no covariation; each speaker made as many conservative as progressive statements. In the condition with high fit, there was a strong covariation, so that members of one category made much more conservative statements than progressive statements and vice versa for members of the other category. The model analyses revealed a main effect of accessibility on the category memory parameters so that category memory increased with accessibility. In addition, category memory also increased monotonically with increasing fit, and there was no interaction between the two factors. Reconstructive category-based guessing was similarly affected by fit in that guessing on the basis of the contents of the statements became more pronounced as fit increased. However, in line with a multiplicative interaction of fit and accessibility, the effect of fit on reconstructive guessing became larger as accessibility increased. Reconstructive category use and category memory both converge to increase overall homogeneity in terms of within-category confusions, especially when fit is high, but the two components appear to be governed by different laws. While accentuation in memory was affected by both accessibility and fit in an additive fashion, the expectancy of homogeneous group attitudes was not only informed by the extent of statement-
116
THE PSYCHOLOGY OF GROUP PERCEPTION
category covariation, but also inter acted with category accessibility: The higher the covariation, the stronger the effects of accessibility on expecting category-wise homogeneous attitudes. RELATIVE GROUP SIZE In a second series of studies, we (Klauer, Wegener, & Ehrenberg, 2002) looked at the effects of numerical group composition. Different theoretical accounts predict that minority groups are generally seen as more homogeneous than majority groups paralleling the in-group versus outgroup asymetry (Mullen, 1991; Simon & Brown, 1987; E.R.Smith & Zárate, 1992). Alternatively, S.E.Taylor (1981) argued that “any person, whether male or female or black or white, is distinctive in inverse proportion to the number of other members of his or her race or sex present in the social context” (p. 94). In her account, comparative distinctiveness is the operating factor and leads to the allocation of a disproportionate amount of attention to the persons that are distinct by virtue of their relative rarity. Thereby, the “more discriminations [are made] within a subgroup, the fewer the members of the subgroup are” (S.E.Taylor et al., 1978, p. 780); that is, person-based processing should increase. Because the distinct persons’ category memberships will also be more salient, increased stereotyping of these persons is also predicted, that is category-based processing should simultaneously increase. Thus, antagonistic components of homogeneity, person memory and category memory, should both be affected by group size. In the second study of the just-mentioned paper (Klauer et al., 2002), young and old men discussed gender roles in romantic relationships. There were three conditions that differed in the numerical composition of the discussion groups. They were composed of either two young men and six old men, four young men and four old men, or six young men and two old men. All discussants made three conservative and three progressive statements. Upon completion of the “Who said what?” task, participants also rated selected discussants on traits that are stereotypically associated with either young or old people. This study thereby allowed the assessment of three different category-based processing modes: category memory, reconstructive use of stereotype-based expectancies in guessing, and explicit stereotyping in ratings. The predictions of the account by comparative distinctiveness received partial support: As expected, both item memory (Hamilton, Dugan, & Trolier, 1985) and person memory decreased as relative subgroup size increased. Category memory, however, was not significantly affected by relative group size in this study. Relative group size also had a significant impact on the other two category-based processing modes, reconstructive guessing and stereotyping in ratings. The likelihood of guessing that a statement was made by one of the members of a given category increased with relative subgroup size. In addition, participants were generally more likely to assign conservative statements to old speakers and progressive statements to young speakers when they were guessing. Interestingly, stereotype-congruent guessing was significantly most pronounced for balanced groups in which there were equal numbers of old and young men and in which relative subgroup size was therefore not a factor that distinguished one group as the likely origin of a given statement whereas stereotype congruency was. Yet another pattern emerged for the rating measure. In line with the account by comparative distinctiveness, stereotyping was maximal for members of the smallest subgroup, that is, for minority members. We concluded that the application of category-based knowledge is governed by the perceived usefulness and functionality of the available information as determined by the particular context of application, whereas memory for stereotype-relevant information is governed by distinctiveness in the encoding context. Relative group size is thus an instance of a variable for which different components of homogeneity dissociate, presumably because they are causally affected by different factors—perceived usability versus distinctiveness—which are in turn affected differently by relative group size.
COMPONENTS OF HOMOGENEITY
117
CROSSED CATEGORIES Crossed categories are given when two usually dichotomous dimensions of categorization are crossed orthogonally. For example, crossing the gender dimension with age might lead to four groups—young men, young women, old men, and old women. Deschamps and Doise (1978) were the first to investigate the consequences of crossing categories for social perception, stereotyping, and discrimination. They found that crossing categories led to reduced stereotyping and discrimination compared to conditions in which only one dimension of categorization was present, although these early findings could not always be replicated (e.g., Crisp & Hewstone, 1999). As argued by Deschamps and Doise (1978), the accentuation effect along each dimension will be counteracted by the accentuation effect along the orthogonal dimension. For example, accentuation along age will lead to increased perceived similarity of members of the same age group. Simultaneously, the accentuation effect along gender will emphasize the differences between the male and female members of a given age group. According to this analysis, crossing categories should lead to reduced group homogeneity. In terms of memory organization, the category memberships on both dimensions of categorization might be laid down as independent traces. Alternatively, subgroup membership might be encoded directly as an integral memory trace. In the former case, subgroup membership would have to be reconstructed from the separate category memory for each dimension of categorization; in the latter case, subgroup membership would be laid down in memory as a meaningful bit of information in its own right. In Table 8.1, the d-parameters for category memory are shown for crossed categories. The four category memory parameters in the body of the table refer to the probabilities with which a particular combination of category memberships on both dimensions of categorization is remembered completely (d12), partially ( and ), or not at all (). In addition, the marginal d-parameters represent category memory for the two categorizations considered individually. They are obtained by summing appropriate d-parameters from the interior of the table; e.g., . If subgroup memory is simply determined by the independent and separate contributions of category memory for the two dimensions of categorization, then the d-parameters in the body of the table should be given by the products of the marginal probabilities. According to the law of parsimony, if this independence model holds, then we do not need to consider the conjunction of both category memberships as a subgroup that constitutes a meaningful entity in its own right. That is, at least with respect to memory organization, subgroup formation need not be postulated to account for such data. Alternatively, if the memory for subgroup membership d12 exceeds the product of the marginal probabilities, i.e. d12>d1. d.2, then this is evidence for the formation of a meaningful subgroup. Stangor, Lynch, Duan, and Glass (1992) and van Twuyver and van Knippenberg (1998) attempted to test for this kind of dependency using assignment errors in the “Who said what?” paradigm. Interactions between the crossed dimensions of categorizations were found and interpreted as evidence for subgroup formation in the organization of memory for the members of the discussion groups. However, assignment errors are not process-pure estimates of category memory, and interactions may be found even in the absence of subgroup formation. For example, person memory, item memory, and guessing processes all affect the assignment errors and contribute to effects found in this dependent variable. We extended Klauer and Wegener’s (1998) model to the case of crossed categories. The extension in particular allowed us to estimate the four category memory parameters in the interior of Table 8.1 directly. Thereby, we could test in a focused manner whether or not subgroup memory is based on the indepen
118
THE PSYCHOLOGY OF GROUP PERCEPTION
TABLE 8.1. Probabilities of Remembering Crossed Categories Category Remembered for Dimension 2?
Category Remembered for Dimension 1?
Yes
No
Yes No Sum
d12
d.2
d1.
1
Sum
dent contributions of category memory for each dimension of categorization or whether both dimensions interact providing evidence for subgroup formation. Similar dependencies between the two dimensions can, however, also arise in reconstructive category guessing. Klauer, Ehrenberg, and Wegener (2003) applied the model to data from four studies that crossed age and gender categories. In each study, the four resulting subgroups consisted of three members each of young women, old women, young men, and old men. It was found that age and gender were used independently in reconstructive guessing, whereas there were strong dependencies in category memory. In fact, category memory appeared to be organized hierarchically so that the age category was encoded only if there was gender memory. That is, participants appeared to be able to remember a person’s age category only if they were also able to remember the person’s gender. In terms of the category memory parameters, the probability of remembering the speaker’s age for a given statement was zero, given that there was no gender memory, that is, =0, and it was significantly larger than zero only when there was gender memory, that is, d12>0. There was, however, independent memory for gender, that is, the probability of remembering the speaker’s gender was substantial even when the speaker’s age could not be retrieved: >0. In terms of homogeneity, this hierarchical organization of category memory implies that frequently category memory will be governed solely by the dominant dimension of categorization. In these cases, perceived group homogeneity caused by category memory will be just as strong as in a situation with single categorization along the dominant dimension, and there will be little intragroup assimilation or accentuation of intergroup differences along the subordinate dimension of categorization. Only if there is category memory for the dominant and the subordinate dimension of categorization is the accentuation effect likely to be reduced along both dimensions simultaneously. Thus, any beneficial effect of crossing categories exerted through the category-memory component of homogeneity is likely to extend to only a relatively small subset of cases. DISCUSSION The “Who said what?” paradigm provides a memory measure of group homogeneity (e.g., Judd & Park, 1988; Lorenzi-Cioldi, 1993; Lorenzi-Cioldi, Eagly, & Stewart, 1995; Ostrom, Carpenter, Sedikides, & Li, 1993). In the extension by Klauer and Wegener (1998), it permits separate assessments of different processes affecting homogeneity; notably, person memory assesses the role of individuating features (Wilder, 1984, 1986), category memory the role of category-level representations of perceived group interactions (Mullen, 1991; Ostrom & Sedikides, 1992; Ostrom et al., 1993), and reconstructive guessing the role of stored beliefs about the extent of group homogeneity (Park & Judd, 1990; Ostrom & Sedikides, 1992; Wilder, 1984) on relevant attitude issues. The paradigm is unique in making these measures available within one and the same unobtrusive task. This renders alternative accounts of possible dissociations
COMPONENTS OF HOMOGENEITY
119
between the different measures in terms of artefacts in measurement procedures such as scale artefacts or order effects unlikely. In fact, a differentiated picture of the processes underlying perceptions of group homogeneity and, more generally, of the processes underlying person and group perception emerges from the studies just sketched. Current theories of impression formation (Brewer, 1988; Fiske & Neuberg, 1990) suggest that person-based and category-based processing should in general correlate negatively. Similarly, social identity theory and self-categorization theory distinguish two mutually exclusive modes of person perception, the intergroup level and the interpersonal level, between which a functional antagonism is assumed to exist (e.g., McGarty, 1999). Yet, in the studies on relative group size, person memory, the measure of person-based encoding, decreased with increasing subgroup size, whereas category memory, the measure of category-based encoding, was not a function of relative subgroup size. In the studies on accessibility and fit, both person memory and category memory tended to increase simultaneously with increasing fit. This latter situation, in which an individuated impression appears to have been formed under a category-based processing mode, has been termed category-based individuation by Brewer, Harasty, & Feinstein (1999). Moreover, there are strong indications that different modes of category-based processing can be meaningfully distinguished and follow different laws. Thus, in the studies on relative group size, three different kinds of categorybased processing were tapped: categorization during encoding as measured by category memory, reconstructive category application during retrieval as measured by the guessing parameters, and stereotyping in ratings. The three measures were affected by relative group size in different ways. In the studies on accessibility and fit, there were additive effects of accessibility and fit on category memory, but reconstructive guessing appeared to follow a multiplicative law combining accessibility and fit. The present work therefore converges with several different lines of recent research in suggesting that functional dissociations exist between different modes of processing, such as stereotype activation and application (e.g., Devine, 1989; Gilbert & Hixon, 1991; Macrae & Bodenhausen, 2000), and different functions of social categorization (e.g., E.T.Higgins, 1996) such as organizing memory and stereotype use in filling memory gaps at a later time. This means that many issues can be expected to require a more elaborated theoretical and empirical analysis than is sometimes proposed. For example, an issue of much debate is the effect of cognitive load on social categorization and stereotyping. When cognitive resources are taxed, people have been argued to remember more stereotypeconsistent than inconsistent information (e.g., Macrae, Hewstone, & Griffiths, 1993), and their judgments exhibit more stereotyping (e.g., Bodenhausen & Lichtenstein, 1987). Recent work suggests, however, that stereotype-inconsistent information can receive special weight under conditions of high cognitive load (e.g., J.W.Sherman & Frost, 2000; Yzerbyt, Coull, & Rocher, 1999). Theoretically, the assumption that cognitive load increases the reliance on category-based processing is an important component of Fiske and Neuberg’s (1990) model, and stereotypes and categories are often seen as simplifying heuristics and energy-saving mental tools (e.g., Macrae, Milne, & Bodenhausen, 1994). This point of view can be labeled the resourcebased view. Spears and Haslam (1997) critically reviewed many of the relevant findings and questioned the assumption of the resource-based view that social categorization increases with increasing load. On the basis of social identity theory and self-categorization theory, these authors argue that social categorization itself is a demanding process. In this view, categorization is based on determining similarities between members of the same category, and building on Oakes’s (1987) work, Nolan, Halam, Spears, and Oakes (1999) call this approach the fit-based view. They argue in particular that cognitive load should decrease categorybased processing, because it impedes the detection and analysis of comparative fit in social
120
THE PSYCHOLOGY OF GROUP PERCEPTION
situations. Fit-based and resource-based views thus lead to opposed predictions for the effects of cognitive load on social categorization. However when different category-based processing modes are distinguished, these theoretical differences and the differences in the reported findings may be found to be more apparent than real. Reconstructive category guessing in particular is likely to depend strongly on the perceiver’s ability to determine whether there is comparative fit in a given context. Thus, reconstructive category guessing is likely to be disrupted when perceivers are placed in a condition with high cognitive load as they observe the original social situation. Reconstructive category guessing would thereby decrease with increasing load. Simultaneously, encoding category membership for individual group members may well follow a different law as a function of cognitive load. In line with the view that categories function as cognitive schemas, perceivers might rely more strongly on accessible categories in encoding a given social situation when they are placed in a condition with high cognitive load. Thus, category memory might increase with increasing load. These considerations as well as the reviewed research argue for a differentiated analysis of different person-based and category-based processing modes. For example, group homogeneity as observed in the assignment errors of the “Who said what?” task can be seen as caused by the interplay of different processes, such as accentuation in memory and beliefs about the extent of homogeneity operationalized, respectively, by the category memory parameters and the parameters of reconstructive guessing. The dissociations that occur between these processing modes may further our understanding of apparently contradictory findings and theoretical positions. The conditions under which dissociations occur should therefore be identified and explained.
9 Forming Stereotypes of Entitative Groups CRAIG McGARTY School of Psychology, The Australian National University
Social psychology has established the consensus over the past 30 years that in order to understand relations between groups we need to investigate the ways in which group members come to perceive and learn about their own and other groups. Indeed, a volume like the present one, which reflects many different perspectives, would make little sense if this consensus did not exist. A common feature of contemporary discussions of the way stereotypes form is their extensive reference to the process of social categorization. This apparent consensus nevertheless conceals some deep divisions within the discipline. Although just about everybody agrees that the process of categorization is used by group members to make sense of the groups with which they interact, it is certainly not true that all scholars agree about the nature of that categorization process. In this chapter I will start by spelling out what I mean by the categorization process and then apply this understanding to stereotype formation. In order to do this I will (a) outline a set of conceptual tools that I have developed for explaining what categorization is, (b) briefly review some relevant research on stereotype formation, (c) explain how the process of stereotype formation can be conveniently described in these terms, and then conclude by (d) considering work on essentialism and entitativity in terms of the same overview. As I discuss in my book Categorization in social psychology (McGarty, 1999) the fault lines in the debates about what categorization is center on the extent to which categorization is either an adaptive, knowledge-creation process that involves effort after meaning or an error-prone process that serves to conserve cognitive resources (see also McGarty, Yzerbyt, & Spears, 2002). These contrasting views of categorization reach their clearest expression in the treatment of social categorization. If categorization is a deficient process, then, it is often argued, social categorization is where these deficiencies are most clearly exposed. Oakes (2000) considers that many treatments of social categorization wrongly equate it with “the root of all evil”. Such negative views of social categorization
122
THE PSYCHOLOGY OF GROUP PERCEPTION
have a long history and they can clearly be seen in the work of writers such as Lippmann (1922) and G.Allport (1954). On the other hand, a number of writers especially, but not exclusively those influenced by selfcategorization theory (Turner et al., 1987a), such as McGarty (1999) and Oakes (Oakes & Turner, 1990; Oakes, Haslam & Turner, 1994), emphasize in contrast the importance of categorization as a meaningcreation or sense-making process. Indeed, as the cognitive psychologist Medin (1988) observed, the principal problem that perceivers face in social perception is not that they have too much information but too little. This alternative stance on categorization, which paradoxically is also clearly present in the work of G.Allport (1954), has been heavily influenced by Bruner (1957) and Tajfel (1981a). The existence of the two stances sets the stage for a debate in the field about what categorization is and about what categorization is for. In terms of a hackneyed old expression, the central debate in categorization involves the question of whether categorization simply allows us to see the forest from the trees (gaining knowledge), or whether seeing the forest prevents us from seeing any trees at all (losing information). This debate can be seen in a variety of areas of cognitive and social psychology (for a review of both fields, see McGarty, 1999; for a review of the social psychological literature, see Spears & Haslam, 1997). In broad terms, however, contemporary cognitive treatments of categorization cannot be characterized in terms of this crude distinction. Categorization is now widely understood to be a process that is geared toward sense making, where information loss is a possible but not necessary consequence (and this depends especially on the particular subprocess of categorization considered: see Gilbert & Hixon, 1991; Klauer & Wegener, 1998; McGarty, 1999). The old idea that categorization is necessarily an error-prone process now appears to be largely discredited (see McGarty, 1999). Rather than raking over the coals of this old debate here, though, I wish to explain a theoretic model that enables us to discuss contemporary and past treatments of categorization. THE CONSTRAINT RELATIONS FORMULATION OF CATEGORIZATION The constraint relations model of categorization is an attempt to provide a descriptive language that allows us to be precise about theoretical constructs. These are theoretical constructs that are ubiquitous in treatments of categorization but are routinely called by many different names. The formulation involves understanding the links between three central elements in the categorization process. These elements are category use (the recognition that entities belong to a category), background knowledge (intuitive theories or models that explain similarities and differences), and perceived equivalence (the perception of similarity and differences between entities). The full range of constraint relations that can exist at a given instant of time is depicted in Figure 9.1. Thus, to take one set of connections at any different time, background knowledge can constrain the use of categories that can in turn constrain the degree of equivalence (similarity) on nonequivalence perceived to exist within and between categories. The reason I use the term “constraint relation” rather than “causal relation” is because the theories that the model seeks to explicate are not themselves always clear about causality. Nevertheless, constraint relations are roughly analogous to “weak causal links between psychological constructs.” For example, background knowledge about similarity and differences may constrain the perception of similarities and differences in the sense that knowing that two things are similar may make us more likely to see the two things as similar in some setting. Thus we can say that background knowledge might constrain perceived
FORMING STEREOTYPES OF ENTITATIVE GROUPS
123
FIGURE 9.1. General form of constraint relations for categorization
equivalence (and by saying this we probably really mean that there is some mechanism of selective activation of knowledge which does have a causal effect on perception that is not being specified). Constraint relations can be illustrated with two analogies. Consider the relationships between climate and weather. It would be imprecise to say that climate causes weather (because in using the term climate we are concealing a variety of causal mechanisms contained within this broad construct), but it would be less imprecise to say that climate constrained weather. To take a psychological example, with respect to selfperception, self-categorization theory argues that long-term commitment to social groups (social identification) can be seen to constrain the salience of self-categorization by increasing the perceiver’s readiness to use that categorization (see Spears, Doosje, & Ellemers, 1997; Turner et al., 1987a). A particular social categorization can be highly salient for both high and low identifiers at some instant, but such instants are more likely at particular times. Returning to the general consideration of categorization processes, some theories do postulate strong and direct causal links. For example, Tajfel and Wilkes (1963) suggested that the explicit use of categories leads to the perception of increased similarities within categories and increased differences between the categories. This hypothesis has been incorporated into a variety of subsequent approaches including selfcategorization theory. Theories of categorization differ in the range of causal links that they specify. Simple constraint relations include those envisaged by Bruner (1957), who argued that fit (related to equivalence) and readiness (related to knowledge) determined category use. These key constraint relations are shown in Figure 9.2. Having introduced this descriptive language for the categorization process, we can now apply it to the particular issue of stereotype formation. The basis of my argument (see McGarty, 2002) is that stereotype formation reduces to the process of developing constraint relations between (a) knowledge about groups, (b) thinking in group terms, and (c) perceptions of similarity and differences within and between that and other groups. Before I articulate the specific constraint relations that seem most relevant for stereotype formation, though, it is helpful to look briefly at the relevant research on stereotype formation. STEREOTYPE FORMATION: EXPECTING AND FINDING DIFFERENCES BETWEEN SOCIAL GROUPS As is well known, treatments of stereotype formation in recent times have focused on the formation of distorted negative stereotypes of minorities. This was not because researchers believed that stereotypes only formed about this sort of group, but evidence for negative minority stereotypes was most consistent with the prevailing metatheory that was coming to dominate social psychology at the time. The theoretical insight at the heart of Hamilton and Gifford’s (1976) attempt to explain stereotype formation was that negative stereotypes of minority groups could be explained by a consistent erroneous association of minority group membership with a distinctive type of behavior. If negative behaviors were particularly distinctive (because
124
THE PSYCHOLOGY OF GROUP PERCEPTION
FIGURE 9.2. Key constraint relations envisaged by Bruner (1957)
they are less socially desirable and therefore more rare than positive events), then this would tend to imply that minorities would be stereotyped more negatively. The importance of this idea for the strictly cognitive approach to social stereotyping was that it implied, while not ruling out the role of social factors in producing consensual stereotypes, that these factors might not be necessary for consensus to occur. The mere fact that different people encountered much the same stimulus information (rare groups and rare behaviors) could result in those different people developing similar stereotypes (as it happens, simulation results obtained by the cognitive psychologists Nosofsky, Palmeri, & McKinley, 1994, suggest this is implausible due to the wide range of different possible categorizations that can be applied by different people—see also S.A.Haslam, 1997). The illusory correlation effect (ICE) in social psychology is that people perceive relations between group membership and behavior when no such relationship exists. In the empirical paradigm that is used to investigate it, subjects are presented with a number of statements (usually 36 or 39) describing the behavior of members of two groups (Group A and Group B). Group B is described as being smaller in the real world than Group A, and, in the stimulus set as a whole, there are twice as many statements about Group A members as about Group B members. Of the statements approximately two-thirds are non-distinctive (normally desirable) and one-third are distinctive (undesirable), but the same ratio of nondistinctive to distinctive behaviors is present in both groups (it is on this basis that any correlation detected by subjects is said to be illusory). A distinctiveness-based illusory correlation would be said to occur where the distinctive behaviors are overattributed to the minority (distinctive) group and members of this group are hence represented less favorably than the majority. In practical terms this means that the minority group is seen as less positive rather than as absolutely negative. The original explanation of the effect was Hamilton and Gifford’s (19976) distinctiveness-based account. The cognitive process underlying ICE is that the co-occurrence of two relatively infrequent events is especially noticeable or distinctive, and this automatically triggers the observers attention. These jointly infrequent events are hence better encoded and more accessible for retrieval. Following the idea of the availability heuristic, the more easily they are retrieved, the more the subject perceives them as numerous, and therefore overestimates their frequency. There is substantial evidence for the illusory correlation effect (for reviews see McGarty & de la Haye, 1997; for meta-analyses see C.Johnson & Mullen, 1993; Mullen & Johnson, 1990) but in the last 10 years a number of new explanations have been developed. The closest of these to the original is the extended distinctiveness-based model of McConnell, Sherman, and Hamilton (1994a). This model holds that illusory correlations need not be based on immediately perceived paired distinctiveness, but that the presentation of
FORMING STEREOTYPES OF ENTITATIVE GROUPS
125
additional stimuli can change the representation of those stimuli so that they become distinctive in the context of the entire set. Thus, distinctiveness is not a constant invariant feature of a stimulus that is fixed at the time of encoding. Rather, previously perceived stimuli can become more or less distinctive as additional comparison stimuli are presented. Greater departures from the original formulation were apparent in the information-loss approaches of Fiedler (1991) and E.R.Smith (1991). The common feature of these independently proposed formulations was that illusory correlation rested not so much on the detection and encoding of distinctiveness-based correlations, but emerged instead from the loss of information. The basis of these explanations is articulated in Fiedler’s (1996) BIAS explanation: where information is lost randomly about groups of different sizes then any clear prevailing impression is likely to be more clearly observed in the larger group. A metaphor is helpful in explaining the basis of this explanation. Imagine two join-the-dots images of the same object, the first of these images is based on 24 dots and the second is based on 12. If we were then to randomly delete half the dots (lose information in the same proportion) from both images our ability to correctly identify the object (after we join the dots) in the first image will be greater than it is in the second case. What this tells us is that impression based on a larger sample of information will be less subject to the vagaries of information loss. In the case of the illusory correlation paradigm both groups are created to have the same prevailing positive nature. What would be expected is that the positive impression of the larger group will suffer less distortion due to random information loss. Thus, the larger group should be seen more positively than the smaller group, without its necessarily being the case that the minority should be perceived negatively, and this is indeed what is generally found. Although very different from the distinctiveness approaches, these approaches nevertheless rest on the role of quite direct (or even no) processing of the stimulus information in producing the effect. A very different approach was taken by McGarty and colleagues (see also Berndsen, Spears, van der Pligt, & McGarty, 2002). These authors propose an approach based on self-categorization theory (a very similar account in terms of evaluative contrast was independently proposed by de la Haye; see McGarty & de la Haye, 1997). The key assumption here was that processing of the stimulus information was powerfully affected by participants’ expectations that there would be differences between groups. The central idea here was the illusory correlation effect derived from routine processes of differentiation between social groups applied to the strange situation created by the researchers. In preliminary research McGarty, Haslam, Turner, and Oakes (1993) discovered that the standard illusory correlation effects occurred under conditions where the participants knew (a) that one group was twice as large as the other, and (b) where they had perceived that there were twice as many positive statements as negative, but where (c) they did not know which statements were associated with which group. This suggested that the effects were not dependent on particular characteristics of the stimulus information used but that the effect emerged from expectations that were created by the experimental setting. That is, participants knew that the statements varied in their desirability, and when they were then asked what the difference between the groups was, they responded in terms of the one dimension that distinguished the statements. As S.A.Haslam, McGarty, and Brown (1996) noted, the conversational logic of this and other illusory correlation paradigms (see Bless, Strack, & Schwarz, 1993) implies that there should be some difference between the groups: why else would the experimenter ask what the difference was? Given that there is a two-to-one ratio of positive to negative statements and a two-to-one ratio of the majority to the minority, then it also makes sense that there should be a connection of the minority group.
126
THE PSYCHOLOGY OF GROUP PERCEPTION
This can be illustrated in the following way. Imagine I were to tell you that there were three behaviors, two of which were positive and one was negative, and that two were performed by Group A and one by Group B. I then ask you to guess whether the Group B behavior was positive or negative. If your answer is guided by base rates, then “positive” is a better guess, but if you assume that I have provided the information about group membership for a reason, namely, that there is in fact some difference between the groups, then “negative” is a good guess. The differentiated-meaning approach to understanding stereotype formation in the illusory correlation paradigm therefore assumes that the experimental situation creates an expectation that there is a difference between the groups and that this motivates a process of interpreting the stimulus information in a way that detects a difference. Participants in the standard illusory correlation paradigm should therefore favor one group over the other (research by de la Haye and her colleagues, see McGarty & de la Haye, 1997, shows that this is the case). The question for the differentiated-meaning account then is why do participants in the standard paradigm favor Group A over Group B? Two answers were provided by McGarty et al. (1993). Both of these rest on the idea that the skewed contingency tables that participants are presented with in the illusory correlation paradigm are open to interpretation as support for the idea that Group A is clearly better than Group B. That is, one way to interpret a skewed contingency table is in terms of a correlation between group membership and desirability, but another way of interpreting such a table is in terms of a contrast between two groups. For example, the stimuli provide 22 pieces of information that Group A performs desirable acts and Group B undesirable but only 18 pieces of information that Group A performs undesirable acts and Group B performs desirable acts. Yet another way to interpret the stimulus information is to see it as providing clear support for the hypothesis that Group A performs more positive than negative behaviors (18 vs. 8), but the evidence for the same hypothesis about Group B is far less clear-cut (9 vs. 4). Clearly this argument is highly similar to that contained in the information-loss accounts of Fiedler (1991, 1996), but the difference is that the differentiated-meaning account invokes the idea of comparison between groups. If each group were considered in isolation, then no group should be perceived more favorably than the other. This aspect of the differentiated-meaning account can also be understood in terms of the principle of diagnosticity (McGarty, Haslam, Hutchinson, & Grace, 1995). Diagnosticity is the degree to which some stimulus pattern is interpretable. This idea has consequences for the concept of entitativity. Originally, D.T.Campbell (1958) used entitativity to refer to the degree to which some group is a real entity. In social psychological usages it has come to refer to the degree to which some group or category is understood to be a real entity (though this concept might be named “perceived entitativity”). Research on categorization tells us that the features that contribute to this perceived reality are the coherence, separableness, and concomitant intelligibility of the category. There is a one-way street between diagnosticity and entitativity in that a collection of things cannot be understood to be an entity unless those things appear in sufficient numbers for it to be safe to conclude that they are to be considered as an entity, in the same way that many people being sick can be identified as an epidemic. Thus diagnosticity obeys the law of large numbers. The metaphor of seeing the forest from the trees is quite apt here. Imagine looking down at the ground from a plane at an altitude of 10,000 meters. At that height you can see forests, but you will fail to detect an equal number of randomly scattered individual trees. Moreover, you will not be able to see a forest unless it is made up of a sufficiently large number of trees. The same point applies in the everyday and in the scientific analysis of social phenomena. It is easy to detect the presence of group processes such as the ingroup bias effect because of the statistical regularity
FORMING STEREOTYPES OF ENTITATIVE GROUPS
127
provided by the norms of the group. The same number of individuals favoring themselves over others is more difficult to observe. Exactly the same statistical processes can apply in the scientific practice of psychology. Imagine a paper with two studies, each with the same effect size, but where one has 26 participants and the other has 13 participants, and the effect was only significant in the first study. Significance tests tell us that we can be confident about the tendency shown in the first case but less confident in the second case. One of the arguments that has been developed in our elaborations of the work of Fiedler and Smith (1991) is that participants in the illusory correlation paradigm may be engaged in remarkably similar decision processes. They are in a position to make the clear inferences that members of Group A perform more desirable than undesirable behaviors (due to the greater diagnosticity of that group), but they can make no such conclusion about Group B. We believe that this difference in diagnosticity provides the basis for the greater perceived entitativity of Group A which faciliates the development of differentiated meaning in the illusory correlation paradigm that the groups are different. As discussed by McGarty et al. (1995), an important contribution to diagnosticity comes in the form of diversity, the richness or variety in the range of positions considered. Unlike unbiased measures of group variability, diversity is sensitive to group size in many more settings. One person can occupy only one position, on a scale at a time, two people can potentially occupy two positions and so on. A higher concentration of persons to positions than is expected can thus be extremely diagnostic of the actual level of entitativity within a group. Thus, a small group that occupies a few positions may or may not be entitative, other things being equal, however, the diagnosticity afforded by a large group occupying a similarly small number of positions can be powerfully informative about that large groups entitativity. That is, other things being equal, low diversity relative to group size will lead to increased potential for entitativity. A caveat is important here. Like similarity, diversity needs to be defined in terms of available dimensions of comparison (along the lines of what Medin, Goldstone, & Gentner, 1993, refer to as “respects” for similarity). Perhaps the easiest way to understand the current view of similarity is that rather than being a mere reflection of stimulus attributes, it is itself an attribution about or interpretation of relations between objects. The diagnosticity of stimulus information can therefore contribute greatly to the potential for perceivers to form the impression that a group is entitative. That is, where the information is not diagnostic, or where the diagnostic information suggests an alternative impression, then perceived entitativity will be low. To recapitulate the above, the differentiated-meaning account has a number of important implications. First, if the effect rests on expectations that there should be differences between groups, then when those expectations are removed, there should be no effects. This was tested by S.A.Haslam, McGarty, et al. (1996), Berndsen, Spears, van der Pligt, and McGarty (1999), Yzerbyt, Rocher, McGarty, and Haslam (1997; see Yzerbyt & Rocher, 2002) and Klauer and Meiser (2000). Haslam et al. found that when participants were told that the majority was comprised of right-handed people and the minority of lefthanded people, there was no illusory correlation effect. Similarly, Berndsen et al. found that when participants were told that the two groups were actually psychology students drawn from different years, there was no effect. In perhaps the most definitive test, Yzerbyt et al. (1997) showed that the perceived validity of the classification had very strong effects. When the classification of the behaviors was ostensibly performed by clinical psychologists (a classification that the participants saw as having a valid basis), then the effect was found to be stronger than when the classification was perceived to be invalid (where the classification of behaviors was attributed to decisions made by a computer program). Finally, Klauer and Meiser (2000) found that when the stimuli varied on gender rather than desirability, there was no effect. If we accept that
128
THE PSYCHOLOGY OF GROUP PERCEPTION
people have good reason to expect evaluative differences between groups, but not gender differences, then this provides good support for the differentiated-meaning account. Secondly, there should be no illusory correlation effects when the groups are presented in isolation. In contrast, the information loss account of Fiedler (1991, 1996) suggests that the effect should apply regardless of whether each group is presented alone or in the presence of another group. In other words, the distinctiveness account and the differentiated-meaning account both predict illusory correlation as a withinsubjects effect (ratings of Group A and Group B by the same person), whereas the information-loss account predicts it as a between-subjects effect (ratings of Group A and Group B by participants in different conditions). This picture was qualified by Fiedler (1991) in his discussion of the unpublished research of Crawley and Regan, who examined exactly this point. Fiedler argued that when perceiving the minority group in isolation the cognitive load on the participants was very low, meaning that it was unusually easy to detect the positivity of the minority group. More recently, however, McGarty and Groves (2000) investigated the illusory correlation effect in a between- and within-subjects design using set sizes varying from 12 to 36 statements in the groups presented alone. They found no evidence that the positivity of groups varied with set size. This is evidence that, although the effect must depend on information loss, the information loss account is not a sufficient explanation. In other words, the differentiated meaning account which incorporates the idea of information loss appears more plausible. Thirdly, the differentiated-meaning account of illusory correlation holds that the effect should be accompanied by attempts to differentiate groups. Research reported by Berndsen, McGarty, van der Pligt, and Spears (2001) using a think-aloud technique lends support to this idea. Participants’ descriptions of what they were thinking about while watching the stimuli show that they were developing and testing hypotheses that there were differences between the groups. That is, when participants were asked to describe what they were thinking about when viewing the stimuli, they expressed that they were trying to work out what the point of the task was, or what the relationship between group membership and desirability of relations was. Another crucial difference was the strategy used by participants to test the hypotheses that they had formed. Participants who had not displayed an illusory correlation generally had adopted the strategy of searching for disconfirming information, whereas participants who did show an illusory correlation effect had tended to focus on confirming information. One further intriguing possibility not explored empirically by Berndsen et al. (see also Berndsen et al., 2002) is that people may be more willing to use a confirmatory strategy for hypotheses they develop themselves and a disconfirmatory strategy for hypotheses given to them by other people such as the experimenter. In other words, where participants are told to expect that the groups will be different on the evaluative dimension, the effect may be undermined. It is even possible that telling participants that there was no difference (difficult to disconfirm) and there was a difference (easy to disconfirm) would produce the same effect of reduced illusory correlation. This evidence suggests that the illusory correlation effect is produced by a dynamic process of differentiation. Further strong evidence for this was obtained by Berndsen et al. (2001, Study 1) who showed that the stimulus information presented in the standard paradigm was reinterpreted in line with the developing expectations of intergroup difference. That is, the same stimuli were rated more positively when they were presented as coming from the majority than the minority. In short, stimuli do not provide stable invariant information about desirability and undesirability; rather, desirability and undesirability are themselves products of the way relations between groups are interpreted. This result undermines the possibility that the effect is produced by distinctiveness of a small set of distinctive stimuli or the mere loss of information about stimuli. Rather, a dynamic and contrastive process of reinterpretation appears to be
FORMING STEREOTYPES OF ENTITATIVE GROUPS
129
occurring during the presentation of the stimuli (and, in broad terms, this is consistent with the emphasis by McConnell et al., 1994a, on changes that occur in perceptions over the course of the presentation of the stimuli). There also appears to be considerable evidence that the effect is structured by expectations of differences between groups. Berndsen, Spears, McGarty, and van der Pligt (1998) provided evidence that the illusory correlation effect was stronger when participants were given the expectation that the groups would be coherent, but that the effect was undermined when there was high variability in desirability of behaviors within groups. Both of these effects are highly consistent with the differentiated-meaning account. When people expect to be presented with coherent groups, they tend to form coherent stereotypes of them. When these expectations are disconfirmed, however, by wide variability within the groups, then stereotypes are not formed. These results show the strong interactive effects of stimulus information (reality) and expectations about that reality in contributing in equal measure to those stereotypes. One piece of evidence that appears more difficult to reconcile with the differentiated meaning account is that the illusory correlation effect appears to be weaker when participants are given impression formation rather than memory instructions (e.g., Pryor, 1986). That is, where participants are instructed to do one of the things that the differentiated-meaning account suggests they are doing all the time (forming impressions), then the effect is weaker than when no such instructions are given. I believe there are two answers to this (though Spears, 2002, offers a more detailed discussion of related points). The first is very simple: Explicit impression formation instructions may make it easier for participants to disconfirm hypotheses about group differences from the outset of the paradigm (in much the same way that they can disconfirm explicit hypotheses given to them by the experimenter). The second answer is more complex and is based on an assumption. I present it here as an intriguing possibility that merits further research. Explicit impression formation instructions in this paradigm also appear to have the somewhat counterintuitive effect of improving memory for category membership at least when compared to memory instructions (though, see chapter 11 of McGarty 1999 for a discussion of similar categorical effects). I can only make this assumption tentatively because there are clear limits on the value of research on correct memory in the illusory correlation paradigm conducted prior to application of the multinomial model by Klauer and Meiser (2000). Accepting this assertion for the sake of argument, though, it is also apparent that there can be no illusory correlation effect in the standard paradigm without substantial loss of information (Fiedler, 1991), so this improved memory may prevent the effect from being observed. Despite the apparent relatedness of impression formation instructions to the process specified by the differentiated-meaning account, explicit impression formation may create a third factor that suppresses the effect. After all, the differentiated-meaning account is proffered as an explanation of the illusory correlation effect rather than of third factors that prevent the effect occurring. On balance, a great deal of evidence supports the differentiated-meaning account, though the evidence certainly does not rule out an explanatory role for distinctiveness or for information loss. One important aspect of the differentiated-meaning account is that it returns discussion of the illusory correlation effect to the issue of stereotype formation. This was the original motivation of Hamilton and Gifford (1976) although their work was essentially restricted to the discussion of the formation of stereotypes of minorities. Subsequent accounts tended to focus concern on the process of illusory correlation in ways that did not appear to be closely related to stereotype formation, but that might provide a clear account of the results of the experimental paradigm introduced by Hamilton and Gifford. The differentiated-meaning account, on the other hand, explains stereotype formation of minorities and majorities in terms of the process of differentiation between social groups that has been so influential in explaining other aspects of intergroup relations.
130
THE PSYCHOLOGY OF GROUP PERCEPTION
FIGURE 9.3. Constraint relations involved in stereotype formation
CONSTRAINT RELATIONS FOR STEREOTYPE FORMATION McGarty (1999) spells out a set of constraint relations for group variability judgments and related phenomena. Much more specific patterns are shown in that book for actual group variability judgments, but the general set covers stereotype formation quite well. The general set of constraint relationships shown by McGarty, however, representd the process as a continuous causal chain. In Figure 9.3 I have laid these out in a linear sequence (which becomes a cycle again by adding feedback loops from Box 6 to Box 1). This sequence begins with the roles of background knowledge and perceived equivalence constraining category use. Where there is background knowledge in the form of expectations that groups should be different and that the groups considered have a real, essential nature, combined with the perception that the information is sufficiently diagnostic to be informative about group differences, then perceivers will be more likely to apply category labels to some situation. This category use leads to the activation of dimensions that are known or believed to differentiate between the groups considered. To the extent that this process is successful (i.e., the right dimensions are found to differentiate between the groups and to take account of differences within the groups) the groups will be seen to be coherent and entitative. These perceptions of entitativity will then be enlisted in updating background knowledge, so that whereas previously the groups were expected to be different, they are now known to be different. This difference is most clear in comparing Box 1 with Box 5. Box 1 relates to the expectations that perceivers bring to the perceptions of a group at a particular time whereas Box 5 relates to the resulting perception. These constraint relations are intended to apply to stereotype formation in general, but the research on the illusory correlation paradigm considered so far illustrates them well. Evidence for the preconditions listed in Box 1 comes from a variety of sources. Research by S.A.Haslam et al. (1996), Berndsen et al. (2001), Klauer and Meiser (2000) and Yzerbyt, Rocher, and Schadron (1997) points to the importance of expectations of differences between groups as preconditions for stereotype formation. Research by Berndsen et al. (1998) shows the importance of expectations of coherence. The final element, essential
FORMING STEREOTYPES OF ENTITATIVE GROUPS
131
qualities, has been investigated by Yzerbyt and Rocher (2002), and its importance is assumed by extrapolation from other domains that are considered in the next section. The elements shown in Box 2 come from evidence that the illusory correlation effect does not occur with very small groups. It is also consistent with evidence obtained by McGarty et al. (1995) that groups were more likely to be perceived to be coherent when there was a sufficiently large sample (though this research does not show the mediating role of Boxes 3 and 4, as these were constrained by the experimental situation). Boxes 3 and 4 are intimately related. The clearest evidence for the process comes from work by Berndsen et al. (2001), which shows that perceivers in the illusory correlation paradigm constantly report thinking in categorical terms and attempts to find ways in which the groups differ. Moreover, when no appropriate dimensions are available for differentiation between the groups (as in the case of the study by Klauer & Meiser, 2000), there is no evidence of illusory correlation. Box 5 is supported by research by McGarty et al. (1995) and by Berndsen et al. (1998). Somewhat contrary results, however, were obtained by McConnell et al. (1997), although they are reconciled by Berndsen et al.’s second study. Finally, evidence for Box 6 comes from a variety of studies by Berndsen and colleagues on reinterpretation of stimulus information (Berndsen et al., 1998; Berndsen et al., 2001). As these expectations become updated, though, it should be expected that they will be constrained further by through processes of social interaction within groups. S.A.Haslam and colleagues have discussed these processes in relation to the topic of consensualization (S.A.Haslam, 1997; S.A.Haslam, Oakes, McGarty, Turner, Reynolds, & Eggins, 1996; S.A.Haslam, Oakes, Turner, McGarty, & Reynolds, 1998; Haslam, Turner, Oakes, Reynolds, & Doosje, 2002), whereby group members develop consensual views of other groups. Obviously, perceivers want expectations that are accurate, and one key way of being accurate is to have expectations that are consistent with the expectations of other people with whom we expect to agree. Indeed, the available evidence suggests that people become uncertain, that is, they doubt that they are correct, when they experience disagreement with relevant others (McGarty, Turner, Oakes, & Haslam, 1993) and become more certain (think they are correct) when they experience agreement (Bliuc, McGarty, & Reynolds, 2000). A fuller treatment of consensualization, however, is beyond the scope of this chapter (see Haslam et al., 2002). What I do propose to discuss in more detail, though, is the key element in Box 1 that has not been discussed so far. My reasons for doing this become clear when we ask a simple question. Are the relations shown in Boxes 1 to 6 expected to apply just as much to the formation of impressions of nonsocial categories as to stereotypes of social categories? The answer I would give is that, although the data I have discussed here focuses on categories of people, all of the principles are expected to apply to both types of categories. Having said that, there is more to the content of social stereotypes, and it would be remiss of me in a book such as this one to neglect these extra values and nuances that come when we form impressions of categories made up of our fellow human beings. The next section therefore supplements the perspective advanced so far to address the distinctive qualities of categories of intelligent, intentional beings. NATURALNESS, ENTITATIVITY, AND COLLECTIVE AUALITIES In recent times a number of authors have argued that the dimensions of entitativity provide an important basis for distinguishing social groups. This indeed provides an important inspiration for the current book. Of the several relevant approaches I wish to focus on the work by Hamilton and colleagues and by (Nick) Haslam and colleagues.
132
THE PSYCHOLOGY OF GROUP PERCEPTION
In both of these lines of research the procedures have involved ratings of different groups and categories on multiple dimensions. The motivation for the work by Haslam, Rothschild, and Ernst (2000) has been to identify the underlying dimensions for distinguishing social categories. The motivation for Hamilton and colleagues’ work has been to identify underlying dimensions and also to establish a typology of groups. Haslam and colleagues have identified that the degree of naturalness and entitativity are different things. Some groups and social categories are defined in terms of the degree that there is believed to be some natural basis for defining them. This is not to say that the categories necessarily have these qualities. All that need be true is that they are believed to have these qualities. Race provides a good example of a social categorization that usually fits this bill. The other key dimension for distinguishing between social categories is in terms of the degree to which they are coherent and exclusively defined (entitative). On the other hand, Hamilton, Sherman, and Lickel (1998) argued that there is an entitativity continuum that distinguishes groups and categories. This dimension combines aspects of the entitativity and essence dimensions discussed by Haslam and colleagues so that the extent to which groups represents natural kinds contributes to perceptions of entitativity. A key emphasis in the work by Hamilton and colleagues that is less apparent in Haslam’s et al.’s work is that Hamilton and colleagues emphasize the perceived group organization, shared goals, and social interaction between the members. Lickel et al. (2000) identified a number of different types of group in research conducted in the United States and Poland. One classification that emerged from sorting tasks included intimacy groups (such as members of a family), task groups (such as an airline flight crew), social categories (such as citizens of a country), and loose associations (such as people at a bus stop). There is much merit to these ideas, but research by Khalaf and McGarty (2002) on perceptions of social categories suggests that while Hamilton and colleagues may have adequately sampled dimensions for distinguishing between social categories and groups, they may have excluded important types of groups and categories. On the other hand, Haslam and colleagues may have sampled too narrowly from the dimensions that underlie group perceptions. This can be seen most clearly in Lickel et al.’s (2000) definition of social categories. Khalaf and McGarty’s study (2002) included a range of different types of categories of people. These included mental (depressed, anxious, and schizophrenic) and medical (diabetic and paraplegic) illness categories, psychological categories (extravert and homosexual), opinion-based categories (Catholic and supporters of a minor nationalist party), ethnic categories (Aboriginal Australian and American), and activity-based social categories (burglars and athletes). The results showed that the most powerful discriminators between the categories involve collective qualities. People distinguished between the type classes primarily in terms of whether the members had a sense of their own superiority and followed organized rules rather than in terms of the dimensions of naturalness or entitativity identified by Haslam and colleagues. In other words, the primary basis for differentiating between categories of people appeared to be the extent to which emergent group formation had occurred in those groups. Extraverts and schizophrenics may be perceived to have relatively stable and consistent behaviors (if they did not, we would not need the categories), but the basis of the coherence of these categories is not social organization. On the other hand, perceived coherence within categories such as a religious or political group is presumably almost entirely due to the attribution of social organization. This suggests to us that Lickel et al. (2000) may have correctly emphasized some broad differences among types of groups, but they may have underspecified important differences within what they term social categories. Their research shows that social categories only have moderate levels of what we term collective qualities, but our research shows that some social categories are utterly saturated with this sort of meaning whereas others are not.
FORMING STEREOTYPES OF ENTITATIVE GROUPS
133
Similarly, although Haslam’s dimensions may not be entirely useful for distinguishing between classes of social categories our results show that they do powerfully distinguish within these classes. For example, schizophrenia appears to have a more strongly essential nature and to be more entitative than depression and anxiety. It is important to bear in mind that in constraint relations terms all of these arguments apply to knowledge about these categories. Following self-categori zation theory I would argue that actual perceptions of social categories are powerfully constrained by the context. That is, whether a group is perceived to be entitative or not at any given time will be constrained by prior expectations about its coherence, but will also vary with the current reality as reflected in the stimulus information. CONCLUSION In this chapter I have developed the argument that the constraint relations formulation is a useful way of understanding the role of the categorization process in the perception of groups and social categories. This is especially true in relation to discussions of group entitativity and variability because this formulation was designed to help understand relations between knowledge and perception in categorization. To put the argument in the strongest possible terms, the discussion of entitativity and related issues reduce purely and simply to a discussion of the role of knowledge about groups and categories on perception. Two key examples of this were discussed in detail in this chapter. The first example related to a large literature on the illusory correlation effect that has tested the differentiated meaning account. The simple version of the message of this research is that stereotype formation is based on a process of differentiation between groups that involves two central ingredients: expectations of differences of groups based on prior knowledge and current interpretations of the stimulus situation and the data represented by the current stimulus information (for a detailed argument in these terms, see also Berndsen & Spears, 1997). The research shows that where participants do not expect differences between groups, the illusory correlation effect does not occur, but when the groups are expected to be coherent, the effect is stronger. Similarly, when the expectations are present but the stimulus information is not readily interpretable in terms of intergroup differences, then the effect is reduced. A reasonably comprehensive set of relations describing the hypothesized processes are presented in Figure 9.3, which includes the idea that expectations are updated following experience. In the last section I focused on the ways in which these analyses of stereotype formation need to be supplemented in order to deal with social categories. Within this I addressed contemporary concerns about the contributions of perceived naturalness of social categories. I suggested, however, that there was a bigger issue that applied to the formation of stereotypes of social categories and this related to the role of group formation in stereotype formation. Thus, it appears that entitativity and naturalness are vital bases for distinguishing between social categories, but that these principles need to be supplemented by a consideration of the collective qualities that underpin social psychological group formation. This is in line with the recent emphases of Hamilton and colleagues, who have argued that social interaction and goals are important determinants of entitativity. The typology of groups advocated by Hamilton and colleagues appears to underemphasize the diversity contained within social categories in terms of the degree to which they are underpinned by group processes. Our results also point to important ways to extend self-categorization theory. Self-categorization theory has emphasized the integral role of the depersonalization of self-perception in ingroup formation. The results
134
THE PSYCHOLOGY OF GROUP PERCEPTION
presented here suggest that the attribution of the process of group formation to outgroup categories works in much the same way. A hallmark of the SCT perspective is that perceivers develop functional and adaptive views of other groups. If other groups are not static but are dynamic systems, then successful perceivers need to be able to predict the behaviors of these groups accurately. If self-categorization theory is correct, then variable selfperception should play an important role in group life, and the ability of perceivers to accurately perceive and predict and survive group life should rest on their ability to detect collective qualities in other groups. A series of isolated, crazed gunmen represent significant dangers to the safety of those near them, but an organized group of terrorists with the same weapons represent a much more significant threat to the society in general. To understand the difference between these things we need to be able to impute group memberships to outgroups accurately. As I have argued elsewhere (McGarty, 2002), it is the ability to do this that breathes group life into social categories.
10 From Basketball to Business Expertise, Implicit Covariation, and Social Judgment PATRICIA W.LINVILLE GREGORY W.FISCHER Duke University
Fred is basketball fan. When he is told that Chris D. plays basketball for one of the leading college basketball teams, Fred makes several inferences about Chris: • He is confident that Chris is young, male, athletic, and tall. • He believes that there is a slight chance (5%) that Chris is less than 6 tall, a modest chance (20%) that he is between 6'0 and 6'3", a large chance (50%) that he is between 6'4" and 6'8", a modest chance (20%) that he is between 6'9" and 7'0", and a slight chance (5%) that he is taller than 7'0". • Finally, when told that Chris is actually 6'11", Fred infers that Chris is probably a good rebounder, but not a very good outside shooter. This example illustrates the theme of this chapter—that mental representations of social categories (e.g., college basketball players) allow inferences not only about the prototypic features of the category (Chris is young, male, tall) or about the variability within a category (the heights of basketball players), but also about the covariation among features describing category members (Chris is very tall so he probably is a good rebounder).1 A great deal of research has investigated how people form mental representations of social category prototypes and the variability within categories. Increasingly, research also addresses how these representations affect other forms of social judgment. Our current work focuses on two questions regarding a less studied aspect of categories— implicit perceptions of covariation among the features of group members. First, how do familiarity and knowledge of a category affect perceptions of the covariation among the features describing category members? Do experts perceive stronger or weaker correlations than novices? For example, are basketball experts more or less likely to perceive that shooting ability covaries positively with rebounding ability? Second, how do implicit perceptions of feature covariation affect evaluative judgments about individual group members? In addressing these questions, we consider not only
136
THE PSYCHOLOGY OF GROUP PERCEPTION
perceptions of broad social groups based on age, race, and profession, but also perceptions of small, highly interactive task-oriented teams. PERCEIVED VARIATION AND COVARIATION AMONG FEATURES OF SOCIAL GROUP MEMBERS Perceptions of Within-Group Variability During the past 20 years, much of the research on intergroup perception has focused on group variability; for example, variability in the “intelligence” of physicists or the “kindness” of the elderly. The typical finding is an outgroup homogeneity effect—people perceive their outgroups to be less variable along individual attributes than their ingroups. This is a robust finding across a variety of naturally occurring groups, including profession, nationality, race, religion, age, college major, and sorority-fraternity affiliation (for reviews, see Brewer, 1993; Linville, 1998; Linville & Fischer, 1998; Ostrom & Sedikides, 1992; Park, Judd, & Ryan, 1991; Sedikides & Ostrom, 1993). A reverse ingroup arises in some cases where people identify strongly with their ingroup, especially with respect to attributes that are relevant to group identity (R.Brown & Wootton-Millward, 1993; C.Kelly, 1989; Mullen & Hu, 1989; Simon, 1992a). In addition, perceived group variability has a strong impact on many aspects of stereotyping and intergroup perception (for a review, see Linville & Fischer, 1998). For example, when people perceive a group to be homogeneous, they are more likely to generalize from an individual to the group as a whole (Nisbett, Krantz, Jepson, & Kunda, 1983; Quattrone & Jones, 1980), and from the group stereotype to individual members (Krueger & Rothbart, 1988; Ryan, Judd, & Park, 1996). But they are less likely to generalize counterstereotypic information about an individual to the group as a whole (Park & Hastie, 1987; Ryan et al., 1996), or to classify an atypical individual as a group member (A.J.Lambert, 1995; Park & Hastie, 1987). Perceptions of Feature Covariation Although research on intergroup perception has mainly emphasized perceptions of group prototypes and single-feature variability, categorization processes also depend strongly on perceptions of covariation among features (Markman, 1999; Smith & Medin, 1981). As Markman noted, creatures with leaves rarely have legs, but creatures with fur almost always do. People are also sensitive to covariation information in social contexts (e.g., Chapman & Chapman, 1967; Crocker, 1981; Hamilton, 1981; Kelley, 1967; Lewicki, 1986; D.J.Schneider, 1973). Perceptions of covariation (often implicit) underlie peoples naive implicit personality theories. They provide a basis for inferring attitudes, traits, motives, cause, and attributions from known attitudes, traits, and behavior. Returning to our basketball example, if one learned that Chris D. is a starting freshman forward for a major college basketball team and that he is only 6'5" tall, one might reasonably infer that Chris is very quick and also a good outside shooter. On the other hand, if Chris is 6'11", one might infer that he is a good rebounder, but relatively slow and a poor outside shooter. In making these inferences, the perceiver is relying on perceived covariations between height, quickness, rebounding, and outside shooting ability (i.e., height covaries positively with rebounding ability but negatively with quickness and outside shooting ability). Despite the central role of perceptions of feature covariation in person perception, they have not received much attention in research on intergroup perceptions. We address this gap in our research. We show that implicit perceptions of covariation are systematically related both to familiarity with groups and to ingroup status. We also show that perceptions of feature covariation play an important role in the judgments people
EXPERTISE, IMPLICIT COVARIATION, AND SOCIAL JUDGMENT
137
FIGURE 10.1. Feature patterns implicitly encode perceived covariation.
make about individual group members. Further, we show that the construct of feature covariation is conceptually and statistically distinct from constructs relating to the variance of individual features. However, in a broader sense, feature covariation is related to the degree to which one differentiates or makes distinctions among category members. Thus, our research on perceived covariation has important connections to the literature on perceptions of group homogeneity. CONCEPTUAL FRAMEWORK AND PERCEIVED COVARIATION HYPOTHESES Feature Patterns Implicitly Encode Feature Covariation To see why feature covariation is important to intergroup perception, consider what comes to mind when you think about a social category, such as college basketball players or politicians. In both cases, you may think of a single, most prototypic type of person. But if you have much knowledge at all of the cat egory, you are also likely to think about a variety of different types of category members, each characterized by a set of features—for example, big slow centers who are strong rebounders but poor outside shooters, or small quick guards who are good ball handlers, good outside shooters, but poor rebounders. We assume that mental representations of social categories include multiple clusters or patterns of features, each describing a different individual or subtype of the category (for studies suggesting social group knowledge is organized into types or subtypes, see Brewer & Lui, 1984; Hewstone, 1994; Kunda & Oleson, 1995; Linville, 1982; Park, Ryan, & Judd, 1992; Park, Wolsko, & Judd, 2001; Weber & Crocker, 1983).2 To illustrate this, the top panel of Figure 10.1 shows one persons simple mental representation of types of people in business—fast-tracker, nine-to-fiver, and number cruncher—each with its own pattern of features. Note how the feature patterns implicitly encode knowledge of covariation among the features of different business types. In this example, the features hardworking and analytical covary positively. The second panel in Figure 10.1 shows a simple repre sentation of college students—including the subtypes nerd, party animal, and artist. Here, studious and sociable are negatively correlated features. The third panel of Figure 10.1 shows one persons simple representation of types of African-American students—athlete, activist, and religious. In this example, self-confident and aggressive are positively correlated features. Finally, the bottom panel of Figure 10.1 shows one persons mental representation of three players from a past Duke basketball team. Here height and rebounding ability are positively correlated. In our work, we
138
THE PSYCHOLOGY OF GROUP PERCEPTION
measure implicitly perceived covariation using participants’ descriptions that are similar to those in Figure 10.1. In short, we use the terms perceived covariation and implicit covariation to refer to the correlations among pairs of features in a perceiver’s mental representation of a social category or group. Similarly, we use measures of perceived variability that reflect the variability with respect to each feature in a perceiver’s mental representation of a social category or group. Thus, we are speaking of implicit perceptions of category properties, perceptions that are reflected in the mental representation but that may not be subject to conscious awareness. The Familiarity-Covariation Hypothesis One central question is whether group membership, familiarity, and expertise have any systematic effects on implicit covariation. For example, do college students perceive stronger (or weaker) correlations among the characteristics of young people (their age ingroup) than among those of older people (their age outgroup)? Do basketball experts perceive stronger (or weaker) correlations among the characteristics of college basketball players than do basketball novices? These questions led us to propose the following familiaritycovariation hypothesis: Greater familiarity with a social group (or domain) leads to lower perceived covariation among the features describing group members (or objects in the domain) (Linville, Fischer, & Yoon, 1996). Here the term familiarity has several possible interpretations—for example, greater contact, greater expertise, or ingroup membership may all lead to lower perceived covariation (at least in those cases where people are more familiar with their ingroup than with their outgroup). Consistency Bias Plus Learning. There are several possible mechanisms for a link between familiarity and perceived covariation among features. The most psychologically interesting mechanism is one that we term consistency bias plus learning. It involves three propositions. First, people’s default expectation is that the characteristics of other people will be semantically or evaluatively consistent. This is consistent with research on cognitive consistency theory and implicit personality theory (e.g., Abelson & Rosenberg, 1958; Rosenberg & Sedlak, 1972). Second, objective correlations among the characteristics of people tend to be relatively low. For example, when trained observers watched people interact and rated their characteristics as they behaved, correlations among these characteristics were very weak. The average absolute correlation was .14, and the degree of semantically consistent correlation was only .04 (D’Andrade, 1974). We, too, have found low correlations when people rated their own characteristics (Linville & Fischer, 2003a) as well as in objective data regarding the characteristics of members of the Duke and University of North Carolina men’s basketball teams (Linville, Fischer, Zauberman, & Rona, 2003). In short, peoples default assumption is that the characteristics of other people tend to be relatively strongly correlated, when in fact they are typically weakly correlated. As a consequence, people will tend to overestimate the correlations among the characteristics of other people. This is supported by research on the illusory correlation effect, an extreme case in which people see correlations where none exist (Chapman & Chapman, 1967; Hamilton, 1981; Hamilton & Gifford, 1976), as well as by findings showing that people overestimate the magnitude of weak correlations, especially by overestimating the degree of semantically consistent correlation (Berman & Kenny, 1976; D’Andrade, 1974). Third, we assume that the tendency to overestimate correlations among the characteristics of other people is moderated by familiarity. As people become more familiar with a domain, they are exposed to a greater variety of feature patterns, many of which will violate the default expectation of evaluative consistency. This should weaken the perceived covariation among the features of members of familiar groups, such as
EXPERTISE, IMPLICIT COVARIATION, AND SOCIAL JUDGMENT
139
one’s ingroups. So greater familiarity or expertise with members of a group should lead to lower perceived correlations among the features describing different group members or types of members. In combination with the first two propositions, this final assumption leads directly to our familiarity-covariation hypothesis. Other Possible Mechanisms
Second-hand exemplars. In Linville and Fischer (1993), we proposed a second-hand exemplars account for outgroup homogeneity. We argued that knowledge of social groups is based in part on first-hand exemplars encountered in direct observation or interaction with group members; and in part on second-hand exemplars conveyed in media images, books, or oral descriptions by friends and family members. We also suggested that second-hand exemplars frequently represent widely shared stereotypes about groups, whereas first-hand exemplars more accurately reflect the diversity within social groups. Finally, we argued that second-hand exemplars are likely to constitute a larger proportion of knowledge regarding outgroups than ingroups. This provides a mechanism for outgroup homogeneity. If we also assume that second-hand exemplars tend to be more evaluatively consistent than first-hand exemplars, then this mechanism also provides an explanation for why people perceive greater covariation among the features of members of outgroups (or other unfamiliar groups) than among the features of members of ingroups (or other familiar groups). Greater incentives to make distinctions. The more contact one has with the members of a group, the more useful it is to make distinctions among them. In previous work, we have suggested that because people usually interact more with members of their ingroup they have greater incentives to distinguish among ingroup members (Linville, Salovey, & Fischer, 1986; Linville, Fischer, &: Salovey, 1989). This assumption is consistent with work showing that people use more levels per feature for ingroups than outgroups (Kraus, Ryan, Judd, Hastie, & Park, 1993). It is also consistent with a broader theme in research on adaptive decision making, namely, that people expend more effort on information processing when the benefits of expending the effort are high relative to the effort involved (Payne, Bettman, & Johnson, 1993). The assumption that people have an incentive to make more distinctions among members of familiar groups not only predicts outgroup homogeneity, it also predicts that people will perceive lower covariation among members of familiar groups. Why? If one’s goal is to make distinctions among people, then one is especially likely to encode those traits of a person that are surprising or unique—for example, traits that are evaluatively or semantically inconsistent with other things one knows about the person. In summary, there are several possible mechanisms, each leading to our prediction that people perceive lower covariation among the features of members of familiar groups, including their ingroups. The mechanisms are complementary, not mutually exclusive. In many contexts, each of these mechanisms may affect perceptions of covariation. Measuring Implicit Covariation There are several approaches to measuring perceived covariation. Most previous studies have used explicit measures in which respondents directly estimate the degree of correlation or contingency between explicitly specified variables, such as height and weight (e.g., Clemen, Fischer, & Winkler, 2000; Nisbett & Ross, 1980). In our work, we rely on perceptions of covariation that are implicit in people’s descriptions of various members or types of members in a social group (see Linville et al., 1996). The Avg R and Avg |R| Measures. To measure implicitly perceived covariation, we begin by asking participants to generate various subtypes of a social group (similar to those in Figure 10.1), and to describe each subtype by circling the level of each attribute that best describes that subtype—for example, low to
140
THE PSYCHOLOGY OF GROUP PERCEPTION
high analytical ability, indecisive to decisive, and low to high leadership ability. This approach is illustrated in Figure 10.2, which shows some of the attributes used in describing different types of people in business.3 FIGURE 10.2. Example of feature rating sheet for business subtypes task Business Type Label _______________________ Low analytical skill Communicates poorly Low verbal skill Lazy Indecisive Low leadership Disorganized Low creativity Low self-confidence
12345 12345 12345 12345 12345 12345 12345 12345 12345
High analytical skill Communicates well High verbal skill Hardworking Decisive High leadership Organized High creativity High self-confidence
We use these subtype descriptions to compute two measures of implicit covariation for each participant. The first is the average absolute correlation, Avg |R|, computed by taking the absolute value of the correlation between each pair of attributes (e.g., analytical skill and leadership), then averaging across all pairs of attributes for that participant. This measure reflects degree of correlation per se, positive or negative. The second measure is the average correlation, Avg R, computed by averaging the signed correlations between all pairs of attributes. This measure reflects evaluatively consistent correlation among attributes—the degree to which more positive levels of one attribute are associated with more positive levels of other attributes.4 Both covariation measures are implicit in the sense that they are derived from subtypes created by the participant, rather than from direct judgments of covariation. This implicit approach to measuring perceived covariation has several advantages. First, it reflects a property of one’s knowledge structure guiding social judgments—the correlation among attributes in that structure—rather than one’s meta-beliefs about covariation. Second, people do not have to be consciously aware of attribute correlations. Lewicki (1986) has shown that knowledge of covariation affects other judgments even when people are unable to consciously articulate the covariation. Finally, in cases where people have not thought explicitly about the degree of relationship among various attributes, this implicit task may be more natural than making direct estimates of covariation. EVIDENCE FOR THE FAMILIARITY-COVARIATION HYPOTHESIS We have tested the familiarity-covariation hypothesis across a variety of groups, including both broad social categories such as age, gender, race, and profession, as well as smaller dynamic groups such as members of college basketball teams. We begin with our work on broad social groups and then proceed to basketball teams. Familiarity and Perceived Covariation in Broad Social Categories Age Groups. In most cases, people are more familiar with their ingroups than their outgroups. In such cases, the familiarity-covariation hypothesis suggests an outgroup-covariation hypothesis: In cases where people are more familiar with their ingroup, they tend to perceive greater covariation among the attributes
EXPERTISE, IMPLICIT COVARIATION, AND SOCIAL JUDGMENT
141
FIGURE 10.3. Perceived feature covariation is greater for age outgroups than age ingroups (Linville, Fischer, & Yoon, 1996)
of out group members. We tested this prediction using our implicit approach to measuring covariation. We began by asking young and older participants to generate various types of young and older people (e.g., grandmother, bingo buff, fraternity guy, pre-med), and then to describe each type by circling a level of each attribute that best describes that type (e.g., unsociable to sociable, lazy to motivated). From these we calculated the Avg R and Avg |R| measures. Three consistent findings emerge (Linville et al., 1996). First, as predicted, both young and older participants perceived greater covariation among attributes of their age outgroup (see Figure 10.3). Second, we directly tested a familiarity-covariation hypothesis using four measures of familiarity with an age group: number of acquaintances in the age group, number of group members known well, contact hours per month with group members, and number of names of group members generated in three minutes. As predicted, those more familiar with each age group perceived lower covariation among the features of group members. The outgroup-covariation effect and familiarity-covariation effect were supported for both measures of covariation—Avg |R| reflecting covariation per se and Avg R reflecting evaluative consistency.5 Both effects remained significant when controlling for number of subtypes generated, favorability of attribute ratings, attributes that were age-stereotypic or not, and perceived variability (standard deviation) with respect to individual attributes. Finally, familiarity acted as a mediator of the ingroup-outgroup difference in perceived covariation. When we controlled for familiarity with the age group, the ingroup vs. outgroup difference in perceived covariation became statistically insignificant. In contrast, controlling for perceived single-feature variability did not reduce this ingroup-outgroup difference in perceived covariation. In short, we found support for both an outgroup-covariation effect and a familiarity-covariation effect for age-related ingroups and outgroups. Further, the outgroup-covariation effect appears to be mediated in part by differential familiarity with the ingroup and outgroup. Gender Groups. In our second study of broad social groups, we studied perceived covariation among the features of male and female college students (Linville et al., 1996). Gender groups are interesting for our purposes because, among college students, differences in familiarity with the ingroup and outgroup tend to be small and statistically insignificant (Linville et al., 1989). Consequently, outgroup homogeneity effects tend to be weaker for gender groups than for age groups or nationality, where differences in familiarity are much more pronounced (Linville et al., 1989). Similarly, memory studies based on the “Who said what?” paradigm frequently fail to detect ingroup versus outgroup differences for gender groups (e.g., Jackson & Hymes, 1987; Klauer, Ehrenberg, & Wegener, 2003; Taylor, Fiske, Etcoff, & Ruderman, 1978). By
142
THE PSYCHOLOGY OF GROUP PERCEPTION
FIGURE 10.4. MBAs perceive lower covariation than undergraduates among the features describing people in business (Linville, Fischer, & Yoon, 1996)
analogy, if familiarity differences were a primary cause of differences in perceived covariation among features, then we would expect little if any difference in perceived covariation for one’s gender ingroup versus outgroup. Our study of gender groups produced two noteworthy findings. First, we found no significant gender differences in our familiarity measures—number of male and female friends and number of male and female close friends. Second, we found no ingroup versus outgroup differences in perceived covariation. In short, with no group familiarity difference, there was no group covariation difference. Business Types. Our third study focused on expertise and familiarity regarding a broad occupational group—people in business—with which all our participants identified (Linville et al., 1996). Our participants were undergraduate and masters-level students in business school courses. The undergraduates had relatively little direct experience in business, whereas the MBA students had several years of work experience. They generated various types of people in business (e.g., fast-tracker, nine-to-fiver, entrepreneur, number cruncher, leader), describing each type by circling a level of each attribute that best described that type—for example, disorganized to organized, communicates poorly to well. We used these lists of feature ratings to calculate Avg R and Avg |R| measures for each participant. The results supported our familiarity-covariation hypothesis. First, those with more years of business experience perceived lower covariation among the features of types of business people (r=−0.32, p<.01). Second, because our MBA students had more business experience than the undergraduates, MBA students perceived lower covariation among the features describing types of business people (see Figure 10.4). Both effects remained significant controlling for favorability and number of subtypes. Third, differences in business experience mediated undergraduate-MBA differences in perceived covariation— that is, controlling for work experience eliminated the undergraduate versus MBA difference in perceived covariation among the features of business people. Racial Groups. Our fourth and newest study focuses on race, addressing whether friendship patterns affect perceptions of feature covariation (Linville et al., 2003). White students generated types of AfricanAmerican male students (e.g., athlete, activist, partier, angry at society, inner city, studious type, religious), describing each type by using 7-point scales to select levels of characteristics (e.g., low to high analytical skills, verbal skills, self-confidence, hostility, happiness). Participants also completed a Friendship Survey in which they first listed 15 “people you know personally,” then turned to a separate page and completed additional information about each person—including gender, age, ethnicity, the number of hours spent interacting with this person per month. The most dramatic effect of friendship patterns involved those who listed any African-American friends as compared to those who did not. White students listing African-
EXPERTISE, IMPLICIT COVARIATION, AND SOCIAL JUDGMENT
143
FIGURE 10.5. White students with African-American friends perceive lower covariation among the features describing types of African Americans
American friends perceived lower covariation among features when describing types of African-American students (see Figure 10.5). Also, those spending more hours with African American friends perceived lower feature covariation. These contact-covariation effects occurred for both measures of covariation, and they remained significant when controlling for favorability of attribute ratings, number of subtypes generated, and perceived variability (standard deviation) of individual attributes. Thus, greater contact in terms of friendship patterns is linked with lower perceived covariation among features of African Americans’—a racial outgroup for white participants. How is perceived covariation among features for types of African Americans related to racial attitudes? In another part of this study, participants completed a feeling thermometer for a series of groups including African Americans and white Americans, as well as the Implicit Attitude Test using black and white names (IAT, Greenwald, McGhee, & Schwartz, 1998). For both measures, those displaying greater racial bias perceived higher perceived covariation among features when describing types of African Americans (again controlling for favorability, variability, and number of types generated). Perceptions of Covariation in High-Entitativity Groups Entitativity refers to those properties that make a group of people a real group, including common goals and outcomes, importance of the group to its members, similarity among members, and interaction among members (Abelson, Dasgupta, Park, & Banaji, 1998; Brewer & Harasty, 1996; D.T.Campbell, 1958; Hamilton, Sherman, & Lickel, 1998; Lickel, Hamilton, Lewis, Sherman, Wieczorkowska, & Uhles, 2000; Linville & Fischer, 2003b). Like most research on outgroup homogeneity, our initial research on perceived covariation focused on broad social groups (age, gender, race, occupation). Broadly defined social categories such as these are perceived to be relatively low in entitativity (Lickel et al., 2000). For example, Lickel et al. (2000, Table 1) reported that gender and racial groups were rated only 5.16 and 5.0, respectively, on a 1–9 scale where 1 indicated “not at all a group” and 9 “very much a group.” Professions (teachers, doctors) scored slightly lower (4.7 and 4.4), so we might infer that the profession “business people” also falls in this low range. By contrast, members of a professional sports team scored highest at 8. 3. Thus, our prior results concern social categories that are low to moderate in entitativity. Would our pattern of findings emerge for high entitativity groups as well?
144
THE PSYCHOLOGY OF GROUP PERCEPTION
In our current research, we are studying perceptions of college basketball teams (Linville et al., 2003). Compared to the other groups we have studied, college basketball teams are relatively high in characteristics that Lickel et al. (2000) link to high entitativity—they are dynamic and personally interactive, and their members share common goals and common outcomes. Among the various types of groups that Lickel et al. studied, organized sports teams were perceived to be highest in entitativity (Lickel et al., 2000, Table 1). Our conceptual framework gives no special status to high versus low entitativity groups. Thus, we would expect the familiarity-covariation hypothesis to apply to high entitativity sports teams as well as to low entitativity social categories.6 The Basketball Study. We studied Duke students’ perceptions of members of the Duke and University of North Carolina (UNC) basketball teams. An analysis of a wide range of objective measures of physical characteristics (e.g., height, weight) and player statistics (e.g., rebounds, assists), taken from university web pages and official publications, revealed that the average correlation between measures was only 0.325. The objective correlations were very similar for the Duke and UNC teams. An interesting aspect of studying perceptions of one’s own and an opposing basketball team is that this ingroup—outgroup context is unique in that one is not actually a member (exemplar) of the ingroup. In adapting our implicit covariation measures to this context, one change was necessary. Because basketball teams have very few members, it makes more sense to study perceptions of specific team members rather than perceptions of types of team members. So, in this study, we asked participants to select from a larger list, 6 team members with whom they were familiar then to describe each player by circling a level on 18 basketball-relevant attributes that best described that player (e.g., height, rebounding, shooting, ball handling, defense, and speed, all on 11-point scales). As in our studies of social groups, we used these profiles to calculate both the Avg R and Avg |R| measures of perceived covariation for each participant. Later, participants made a judgment of the overall basketball playing ability of each player. Finally, participants completed measures of familiarity with the Duke and UNC teams, social identification with their basketball team (e.g., number of games attended, how important Duke basketball was to them personally), and objective basketball expertise (e.g., factual questions such as identifying Duke and UNC players and positions, ranking of teams in the ACC, tournament winners). Our hypotheses parallel those for our earlier studies of larger social groups. First, our familiaritycovariation hypothesis predicts that those more familiar with a basketball team will perceive lower covariation among basketball features (e.g., shooting and rebounding). Extending this hypothesis to expertise as well, those with greater basketball expertise will perceive lower covariation. Finally, assuming that Duke students are more familiar with the Duke team than with the UNC team, the outgroup-covariation hypothesis implies that people will perceive lower covariation among basketball features for their own school’s team. (The two teams were roughly comparable in covariation for a set of objective player measures.) Do Greater Familiarity and Expertise Lead to Lower Implicit Covariation? In this study, too, the results supported our familiarity-covariation hypothesis. Students who rated themselves as being more familiar with a basketball team (either Duke or UNC) perceived lower covariation among basketball features for that team (r=−0.30, p<.05). Similarly, those with greater objective expertise regarding a team (either Duke or UNC) also perceived lower covariation among basketball features for that team (r=−0.38, p<.01). These effects remained when controlling for favorability of their attribute ratings and perceived variability (standard deviation) of individual attributes. Although students rated themselves somewhat more familiar with their own school team, they had virtually equal objective expertise regarding players from Duke and UNC. On average, knowledge regarding both teams was high. So, under these conditions, our theoretical framework predicts that ingroup
EXPERTISE, IMPLICIT COVARIATION, AND SOCIAL JUDGMENT
145
vs. outgroup differences in perceived covariation should be small, if present at all. This is what we found. On the average, there were no significant differences in perceptions of covariation among the attributes describing members of the Duke and UNC basketball teams. In addition, we calculated for each participant a single-feature variability measure—the average standard deviation of each participants trait ratings of the 6 team members, averaged over the 18 features. As predicted by our familiarity-variability hypothesis, those more familiar as well as those with greater objective expertise regarding a basketball team (either Duke or UNC) perceived greater single-feature variability among the players on that team (rs=0.45 and 0.46). On average, there were no significant differences in perceptions of variability of the attributes describing members of the Duke and UNC basketball teams. In short, from a theoretical standpoint, our findings for perceptions of athletic teams parallel our earlier findings for broad social categories. Those with greater familiarity perceive lower covariation among features, as predicted by our familiarity-covariation hypothesis. This new research extends our initial findings in several ways. First, it shows that our hypotheses apply to small, high-entitativity groups (in this case, sports teams) as well as to broadly defined social groups based on age, race, or nationality. Second, it shows that our hypotheses apply to objective expertise as well as to self-rated familiarity. Basketball Covariation Accuracy. If novices perceive higher covariation among features than do experts, who is more accurate? Are novices overestimating feature correlations compared to objective data, are experts underestimating correlations, or are both overestimating? In other words, where is the bias— with experts or novices? To begin to address this issue, we examined the accuracy of perceived correlations among basketball features for our basketball experts and novices relative to objective statistics. For some of the features that participants evaluated (e.g., height, rebounding, shooting ability), we can obtain objective statistics for the players on the teams in question, then calculate the actual correlations among these objective performance features. We can then compare this with our participants’ perceived correlations among the pairs of features for which there were objective statistics. Analyses showed that across all participants, these perceiver estimates displayed higher correlations than the objective player performance correlations. So, as our theoretical framework assumes, in general participants overestimated the strength of the relationships between the performance measures (e.g., correlation between shooting ability and rebounding). However, the greater a person’s objective basketball expertise, the less the tendency to overestimate the amount of correlation present (r =−0.44). For those highest in basketball expertise, bias in Avg R converged toward 0, whereas for those lowest in expertise, average bias was on the order of 0.35 (i.e., their perceived correlation among features minus the objective correlation among those features was 0. 35). A similar pattern emerged for familiarity. Greater familiarity was associated with lower bias (r=−0.34). While the Avg R bias was sizable (roughly 0.33) for those least familiar with a team, it converged toward 0 for those most familiar with a team. In short, those less familiar or less expert display a bias toward overestimating feature covariation, whereas experts are relatively accurate and unbiased. Social Identity with One's Basketball Team. Might perceptions of team players be influenced by one’s social identity (e.g., identification with the team, importance of the basketball team to them personally, game attendance, wearing school sports logos on game days)? Our familiarity-covariation model makes no predictions regarding social identity, but our results show an interesting pattern. As one might expect, students who identified more strongly with their school’s basketball team evaluated their team players more favorably (r=0.43). Those with stronger team identity also perceived lower covariation among the features of their team players (r=−0.33). The relationship between strong team identity and lower perceived correlations among players’ features is an interesting finding because social identity is not significantly
146
THE PSYCHOLOGY OF GROUP PERCEPTION
FIGURE 10.6. Greater feature covariation leads to making fewer distinctions and to making more extreme overall evaluative judgments
correlated with either self-rated familiarity or objective expertise with the teams. Finally, social identity does not mediate the effect of familiarity or expertise on perceived covariation. FEATURE COVARIATION AND EXTREMITY OF EVALUATION The Covariation-Extremity Hypothesis We have seen that across a variety of types of social groups, including both low-and high-entitativity groups, greater familiarity and expertise is associated with perceiving lower covariation among features describing group subtypes or members. But why should we care about perceived feature covariation? A variety of studies demonstrate that single-feature variability has an important impact on social judgments about group members (for reviews, see Linville, 1998; Linville & Fischer, 1993, 1998). Do perceptions of feature covariation have similar effects? For example, do people who perceive high covariation among features of basketball players form more (or less) extreme evaluations of individual players? Why would perceived covariation be linked with evaluative extremity? Figure 10.6 provides a simple illustration. Imagine that one views nine members of a group, encoding each member in terms of two attributes—ability and motivation—and makes an overall evaluation of each member. If one perceives high covariation among features, especially in the Avg R sense, then one will tend to perceive members of a group as having evaluatively consistent profiles of features. Thus a new target may be encoded as more consistent—either good in most respects, bad in most respects, or average in most respects. The result is more extreme overall evaluations of members. On the other hand, if one perceives low covariation among features, then it is more likely that a new target will be encoded as
EXPERTISE, IMPLICIT COVARIATION, AND SOCIAL JUDGMENT
147
being good in some respects yet bad in others. The result is more moderate overall evaluations (see Figure 10.6). In the example in Figure 10.6, we assume that the overall evaluation of a person is the average of the separate evaluation of the person’s Ability and Motivation. For example, the evaluation of a person who is high in Ability but low in Motivation would be (3+1)/2=2. The standard deviation of a set of overall evaluations provides a useful measure of Evaluative Extremity. As can be seen in Figure 10.6, this Evaluative Extremity score is substantially higher in the High Covariation example on the left than in the Low Covariation example on the right. Note that in this example, the average evaluation with respect to each feature and the variability of each feature is identical for the High and Low Covariation examples. So in this example, the difference in Evaluative Extremity is due entirely to the difference in covariation among features. This analysis provides the basis for our new covariation-extremity hypothesis: Those perceiving greater covariation among the features describing members of a social group (or domain) will make more extreme overall evaluations of members from that group (or objects in that domain) (Linville et al., 2003). Note that this covariation-extremity hypothesis results from evaluatively consistent correlations, rather than from strong correlations per se. For example, if the integrity of politicians were perceived to be strongly negatively correlated with competence, then those who were perceived as good in one respect would be bad in the other. This would lead to more moderate, not more extreme, overall evaluations. However, because Avg R and Avg |R| tend to be positively correlated in actual data, we would expect to observe the effect for both measures, but it should be somewhat stronger for Avg R. Covariation and Extremity of Evaluations of Basketball Players To test the covariation-extremity hypothesis, we return to our basketball study (Linville et al., 2003). In addition to the task used to measure implicit covariation among basketball features, participants also later rated each specific player— “Overall, how good a college basketball player do you think Player 1 is?” (using a scale ranging from “very poor for the ACC” to “excellent for the ACC”). We calculated evaluative extremity in terms of the standard deviation of the overall evaluative rating for individual players. That is, we define overall extremity of judgment as greater range or variability in overall evaluations of individual basketball players. Our results provided strong support for the covariation-extremity hypothesis. Those who perceived higher correlations among the features describing members of a basketball team made more extreme evaluations (larger standard deviations) of individual team players. This was true both when Duke students rated Duke players and when they rated UNC players (rs=.39 and 42, ps< .01). These results clearly show that perceived covariation among the features of a group has an important impact on judgments about individual members. Additional analysis showed that perceived variability of single features among players also had a positive impact on extremity of overall evaluations of players. Interestingly, perceptions of variability were not significantly correlated with perceptions of covariation for either the ingroup (r=.04) or outgroup (r= −.15). Further, the covariation-extremity effect remained highly significant when controlling for favorability of their attribute ratings and perceived variability of individual attributes. Thus, the covariation-extremity effect was not simply a byproduct of differences in perceived variability. Rather, perceptions of covariation and variability had strong and independent effects on the extremity of overall evaluations of these basketball players.
148
THE PSYCHOLOGY OF GROUP PERCEPTION
Extremity of Evaluations of Age Groups To further examine the link between covariation and evaluative extremity, we conducted another age study in which college students generated subtypes of older adults (Linville et al., 2003). They also completed familiarity tasks regarding their friendships with the elderly. Later they made overall favorability ratings for a series of vignettes describing both younger and older targets that were generally positive (e.g., the target overcame a fear of heights to join friends on a picnic) or generally negative (e.g., the target was awkward and withdrawn at a party). Greater range in overall evaluations of the vignettes describing older targets indicated greater evaluative extremity regarding older people. Our results support a covariation-extremity link. Those perceiving higher correlation among features of older people made more extreme evaluations of individual older targets. Familiarity had an impact—those less familiar with older people perceived higher covariation among the features of elderly, and made more extreme evaluations of individual older targets. Further analyses suggested that perceived covariation is a mediator of this link between familiarity and extreme judgments. We do not mean to suggest, however, that greater expertise is necessarily associated with less extreme evaluative judgments. There are several reasons why this might not be the case. First, experts not only perceive lower covariation among features, which leads to more moderate overall evaluations, but experts also perceive greater variability in individual features, which lead to greater extremity of overall evaluations (Linville et al., 2003). Second, we have evidence that experts may perceive covariation more accurately than non experts. In some contexts, objective correlations among features may be very high. If experts perceive covariation more accurately in these cases too, they are likely to perceive greater covariation. Finally, there is recent evidence indicating that the links between expertise and evaluative extremity can be complex. Brauer, Chambres, Niedenthal, and Chartard-Pannetier (in press) cleverly show that experts whose task is to give clear, unambiguous advice (e.g., political advocates) make more extreme overall evaluations, whereas experts whose task is to convey complex information about different features of alternatives (e.g., political journalists) make more moderate overall evaluations. DISCUSSION AND CONCLUSIONS Main Findings So what do experts about business, basketball, and the elderly have in common? First, we have shown that greater familiarity or expertise regarding a social group leads to lower perceived covariation among the features of group subtypes and members. Our empirical tests of this familiarity-covariation hypothesis have included a variety of types of familiarity—amount of contact or experience with a group, ingroup versus outgroup status, and objective expertise. Our empirical tests of this link have also considered a wide range of groups, including relatively low-entitativity social categories (e.g., age and racial groups) as well as relatively high-entitativity groups (basketball teams). The basic familiarity-covariation link was supported for all. Second, we have shown strong support for the covariation-extremity hypothesis that lower perceived covariation leads to less extreme overall evaluations of individual group members. With lower perceived covariation, a perceiver tends to encode group members as good in some respects yet bad in others, which leads to less extreme overall evaluations. In short, from basketball to business, from age to race, greater familiarity and expertise regarding a social category lead to lower perceived covariation among the features describing group members. In addition, lower perceived covariation leads to less extreme overall evaluations of those members.
EXPERTISE, IMPLICIT COVARIATION, AND SOCIAL JUDGMENT
149
Links to Group Homogeneity and Entitativity How do our findings on group covariation relate to other themes of this volume—group homogeneity and group entitativity? Greater Perceived Covariation Is Related to Making Fewer Distinctions among Group Members. The construct of group homogeneity has many interpretations (see Linville, 1987, for a review). At a broad conceptual level, perceived homogeneity is related to the degree to which one makes distinctions among group members. Some measures of homogeneity explicitly capture the degree to which people make distinctions—for example, number of subtypes within a category (e.g., Linville, 1982; Park et. al., 1992); the H-statistic7 measuring dimensional complexity (Linville, 1982; Linville & Jones, 1980); attribute articulation measuring the number of levels used in representing category members with respect to an attribute (Kraus et al., 1993); and the Pd statistic measuring the probability of differentiating among group members with respect to an attribute (Linville et al., 1989). Our current research on perceived covariation has strong ties to the literature that interprets homogeneity in terms of making distinctions. In general, greater perceived covariation among features is associated with making fewer distinctions among group members. Figure 10.6 illustrates this point. In this example, both perceivers use three levels of each attribute (e.g., low, average, high). The perceiver whose mental representation is depicted in the left-hand panel implicitly sees the three attributes as perfectly positively correlated for members of the category. This perceiver distinguishes only three distinct types of category members: (X=low, Y=low), (X=average, Y=average), and (X=high, Y=high). On the other hand, the perceiver whose mental representation is depicted in the righthand panel perceives the two attributes to be unrelated. This perceiver distinguishes nine distinct types of people: (X=low, Y=low), (X=low, Y=average), (X=low, Y=high), and so forth. In short, the perceiver in the right-hand panel makes many more distinctions than the perceiver on the left. Generalizing this example to moderate correlations as well, perceiving higher covariation among features corresponds to making fewer distinctions. Thus, research on perceived covariation is an important and natural extension of previous research on homogeneity. Feature Covariation Is Conceptually Distinct from Single-Feature Variance. Although it is appealing to conceptualize perceived group homogeneity in terms of making distinctions, there are many interpretations and measures of the homogeneity construct. A large body of research has focused on measures of the perceived variability of group members with respect to individual attributes. The two most common measures here are the standard deviation (e.g., Linville et al., 1989; Park & Judd, 1990) and the range (high-low) of the perceived distribution of group members with respect to a specific attribute (Park & Judd, 1990). The question naturally arises—Is perceived covariation related to single-feature variance (or standard deviation)? There are reasons to think that it might be. Most social scientists are taught that it is difficult to detect correlation be tween a pair of variables if one has insufficient variance on either variable. This suggests that the greater the variance of random variables X and Y, the greater the correlation between them will be. If so, our findings regarding perceived covariation might be accounted for by differences in variability. (Actually, any tendency in this direction works against our hypotheses because we predict that high familiarity leads to greater perceived single-feature variability but lower covariation.) One strong argument against this potential confound comes from our data. All of the statistical results we have reported control for single-feature variance and favorability. Greater familiarity leads to lower perceived covariation and greater perceived covariation leads to greater extremity of overall evaluations, even when controlling for differences in perceived variability.
150
THE PSYCHOLOGY OF GROUP PERCEPTION
A second argument against the suggestion that our effects may be partly due to differences in perceived variability is conceptual in nature. From the standpoint of statistical theory, greater variance with respect to two variables does not necessarily result in a stronger (or weaker) correlation between them. A familiar example of this is the bivariate normal distribution in which the degree of correlation between the two variables is unconstrained by their variances (provided that both variances exceed 0, a necessary property of any random variable). This seems to contradict the intuition that greater variance makes it easier to detect correlations. To reconcile this intuition with the statistical principle that variance and correlation need not be related, it is necessary to distinguish between how much variance is present in the population distribution of a random variable and whether a sample is taken from a restricted range of that variable. Sampling from a restricted range of one (or both) variables can reduce the absolute magnitude of the expected sample correlation between the two (Einhorn, 1982).8 For example, if we wish to assess the correlation between height and weight, but restrict our sample to individuals who are 6 feet tall or taller, the expected correlation is smaller than that in the population as a whole. However, if we sample from the full range of two random variables, there is no a priori link between the variance of either variable and the magnitude of the correlation between the two.9 In all of our empirical studies, for example, our measures of perceived covariation have near-zero correlations with single-feature variance. The numerical example presented in Figure 10.6 illustrates the conceptual independence of covariation from single-feature variability. The scatter plots in the left and right hand panels have identical single features means (M=2.0) and identical single-feature variability (SD=0.82), but differ dramatically in the degree of correlation between the two features (Avg R=1.0 versus Avg R=0.0). To summarize, feature covariation is conceptually distinct from the traditional construct of single-feature variability used in outgroup homogeneity studies, and is usually only weakly correlated with measures of perceived variance.10 Thus, studying perceived covariation in addition to perceptions of singlefeature variance is interesting for much the same reason that studying perceptions of group variability is interesting in addition to studying perceived group means or prototypes. Measures of perceived covariation complement traditional measures of single-feature variability and thereby enrich our understanding of the homogeneity construct as well as inter-group perception. In addition, measures of covariation have distinct effects on other forms of social judgment, for example adding insight into the link between familiarity and extremity of evaluation. As the research described here demonstrates, greater familiarity leads to lower perceived covariation which leads to less extreme evaluative judgments. Links to Group Entitativity. It is less clear how the construct of feature covariation relates to that of group entitativity. The existence of shared goals, common fate, and the other characteristics of high entitativity (Hamilton et al., 1998; Lickel et al., 2000) do not necessarily imply that the features describing group members will be strongly (or weakly) correlated. In relatively high entitativity groups based on interdependence, where members interact in the pursuit of shared purposes yet are specialists serving complementary functions (e.g., a college basketball team or a work team), one might expect weak or even negative correlations among some task-relevant features. On a basketball team, those who are best at handling the ball and running the offense are rarely those who are best at rebounding or inside shooting. Similarly, in an organized work group, those who are best at one skill (e.g., writing computer code) are rarely best at other skills (e.g., organizing teams of people), leading to weak or even negative correlations among these abilities. For example, in dynamic, interdependent work groups, we find a negative relationship between covariation and group entitativity (Linville & Fischer, 2003b). On the other hand, in groups based on feature similarity (e.g., age, race), it is less clear what relationship there may be between entitativity and feature covariation.
EXPERTISE, IMPLICIT COVARIATION, AND SOCIAL JUDGMENT
151
In short, there does not appear to be any necessary connection between perceived entitativity and covariation. Furthermore, our findings suggest that the links between familiarity, expertise, perceived covariation, and evaluative judgment are essentially the same for high and low entitativity groups. Thus, the processes governing perceptions of covariation appear to operate in parallel and largely independently of those affecting perceptions of group entitativity. So although implicit representations of covariation have important ties to traditional issues in intergroup perception, the issues raised are new and the hypotheses distinct. The present research demonstrates some of the potential benefits of expanding intergroup research to include notions of perceived covariation. Across a broad range of social categories and groups, we find that greater familiarity and expertise lead to lower perceived covariation among the features describing group members. In addition, lower perceived covariation leads to less extreme overall evaluations of those members. AUTHOR NOTE Preparation of this chapter was supported in part by National Science Foundation Grant SBR-9709615 to both authors. We thank the participants of the EAESP Small Group Conference on Group Homogeneity and Entitativity, Louvain-la-Neuve, 1999, for their very helpful comments, and especially Marilyn Brewer, Chick Judd, and an anonymous reviewer for their comments on a prior version of this chapter. Please direct correspondence to Patricia W.Linville, Fuqua School of Business, Duke University, Durham, NC 27708. Email:
[email protected]. NOTES 1. We use the term “feature” to refer to levels of both binary attributes such as gender, and multilevel or continuous attributes such as height. 2. Clearly, mental representations of social categories include other types of information as well—e.g., visual images, storylike representations of events, and so forth. For our purposes, it is only essential that the representation also include information regarding the features of different types of category members. 3. In our earliest work, we asked people to circle one of three qualitative levels of an attribute—e.g., lazy to motivated. In current work, we use 1–5, 1–7, 1–9, or 1–11 point scales. Calculating implicit measures of variability and covariation using scales with only three levels appears to provide less discrimination than using scales with more levels. Further, with fewer levels, there is greater potential of confounding of mean, variance, and covariation. Thus we recommend using at least five attribute levels. 4. Several previous studies have used measures similar to ours in some respects yet different in others. Linville and Jones (1980) computed group-level absolute correlations among trait ratings of two hypothetical law school applicants, finding stronger trait correlations for racial outgroup applicants. Judd and Lusk (1984) computed subject-level absolute correlations among ratings of the features of rock-and-roll bands and sororities. Park, Ryan, and Judd (1992) also used a measure of absolute perceived covariation as one component of an index measuring both redundancy of attributes and redundancy of subtypes, finding a nonsignificant tendency to perceive greater redundancy for outgroup members. None of these studies included a measure corresponding to our Avg R measure and thus did not discriminate between covariation per se and evaluatively consistent covariation. 5. Note that Avg |R| cannot be less than Avg R. If all interattribute correlations are positive (or zero), then the two measures are equal. But if any correlations are negative, then Avg |R| must be greater. 6. There is one important caveat. Our assumption that the features describing group members are not highly correlated is supported for relatively large populations and broad sets of features. With a small group of people and a small set of features, however, correlations among features might be high. In this case, our familiaritycovariation hypothesis does not apply.
152
THE PSYCHOLOGY OF GROUP PERCEPTION
7. The H-statistic is measured using a trait-sorting task in which subjects select traits that describe subject-generated subtypes within a group. It is a measure of dimensional complexity of a knowledge structure, reflecting the number of independent binary dimensions underlying a subjects trait sort (Attneave, 1959; Scott, 1969). High values of H occur when participants use numerous traits in a relatively non-redundant (low covariation) fashion. Thus, the Hstatistic is similar in spirit to our present measures of perceived covariation. However, our present measures directly reflect covariation, whereas H reflects covariation only indirectly (i.e., high redundancy results in fewer underlying independent dimensions). Further, H is not sensitive to the distinction between evaluatively consistent and absolute covariation that is reflected in our Avg R and Avg |R| measures. 8. More precisely, this is true only if one samples from a contiguous range of the variable—e.g. from the 50th percentile to the upper extreme. If one samples only from the upper and lower tails—excluding the middle—one can actually increase the expected correlation. 9. One can see why by examining the formula for the sample correlation coefficient. This formula can be rewritten as a function of the sum of the product of the Z-scores for the two variables, as shown below
Here, rxy is the sample correlation coefficient, Sx and Sy denote the sample standard deviations of X and Y, respectively, and and are the standardized scores of observations xi and yi, respectively. That is, a correlation coefficient implicitly standardizes random variables before determining the degree of association between them. Translating each variable to its standardized score effectively neutralizes any possible effect of the standard deviation of each variable (provided that both are greater than 0). For example, if we measure height in inches, instead of feet, the variance of the heights will increase by a factor of 144. However, despite this great increase in the variance of the heights, the Z-scores for height will not be affected. Consequently, the correlation between height and weight will also be unaffected. 10. Discrete random variables are a partial exception to this rule. For example, in a two-way contingency table, the marginal distributions of the random variables constrain the degree of correlation that is possible. However, in many continuous multivariate distributions, such as the bivariate normal, the mean and variance are completely independent.
11 Gender Outgroup Homogeneity The Roles of Differential Familiarity, Gender Differences, and Group Size MARK RUBIN University of Newcastle, Australia MILES HEWSTONE University of Oxford RICHARD J.CRISP University of Birmingham ALBERTO VOCI Università di Padova ZOË RICHARDS Cardiff University Outgroup homogeneity (OH) is the tendency to perceive members of one’s ingroup as being more variable than members of an outgroup. The OH effect has been identified using a variety of different measures (for a review, see Park & Judd, 1990). Measures of perceived group stereotypicality assess the degree to which group members are perceived to possess stereotypic versus counterstereotypic traits. OH is evident if outgroup members are perceived to be more stereotypical of the out-group than in-group members are perceived to be of the ingroup. Measures of perceived intragroup dispersion assess the degree to which group members are perceived to be dispersed around the group’s mean position. OH is evident if outgroup members are perceived to be less dispersed than in-group members. Measures of perceived intragroup similarity assess the degree to which group members are perceived to be similar to one another within their group. OH is evident if outgroup members are rated as being more similar to one another than ingroup members are to one another. Finally, name-matching measures assess intra-and inter-category confusions when attempting to recall matches between individual group members and information to which they have been previously linked (e.g., Lorenzi-Cioldi, Eagly, & Stewart, 1995). OH is evident if more intracategory confusions are made for outgroup members than for ingroup members. The OH effect has been found using a variety of natural social groups, including Australians and Americans (S.A.Haslam, Oakes, Turner, & McGarty, 1995), Hindus and Muslims (Hewstone, Islam, & Judd, 1993), Labour party and Conservative party supporters (Kelly, 1988), young and old people (Linville, Fischer, & Salovey 1989, Experiment 1), and firemen and funeral employees (Codol, 1984; for reviews, see Devos, Comby, & Deschamps, 1996; Linville, 1998; Mullen & Hu, 1989; Ostrom & Sedikides, 1992; Voci, 2000). Given the pervasive nature of the OH effect among natural groups, one might expect to find that the effect also occurs for gender groups, especially considering that gender provides a primary basis for social categorization (Brewer & Lui, 1989; J.A.Williams, 1984). However, there has been some debate concerning the evidence for gender OH. Some researchers have found small but significant gender OH effects (e.g.,
154
THE PSYCHOLOGY OF GROUP PERCEPTION
Park & Judd, 1990), whereas others have found no significant effects (e.g., Linville et al., 1989, Expt. 3). Two explanations have been put forward for these equivocal findings. First, Linville and colleagues (e.g., Linville et al., 1989) proposed that OH is driven by differences in familiarity with ingroup and outgroup members. They argued that strong gender OH effects are unlikely to occur because there is usually very little difference in the familiarity that people have with men and women. Second, Lorenzi-Cioldi and colleagues (e.g., Lorenzi-Cioldi et al., 1995) proposed that gender OH is perceived by men but that the reverse effect of in-group homogeneity (IH) is perceived by women. Consequently, overall gender OH effects (i.e., the average effects shown by male and female participants) are unlikely to occur because the IH shown by female participants subtracts from the OH shown by male participants to leave a null effect overall. Lorenzi-Cioldi et al. attribute these gender differences in gender OH to status differences between genders. In this chapter, we review the evidence supporting each of these two explanations for null gender OH effects. Specifically, we look at the roles of (a) differential intergroup familiarity and (b) gender differences in explaining null gender OH effects. We then go on to consider how group size might influence OH. Finally, we report some of our own research investigating the roles of differential familiarity, gender differences, and group size in determining gender OH, and we conclude by discussing a tentative model of OH. EXPLANATIONS FOR NULL GENDER OH EFFECTS Differential Intergroup Familiarity Linville and colleagues (Linville, 1998; Linville & Fischer, 1993; Linville et al., 1989; Linville & Jones, 1980; Linville, Salovey, & Fischer, 1986) proposed that in most cases people are more familiar with ingroup members than with outgroup members. This differential intergroup familiarity may cause OH for three reasons. First, people have a more diverse and articulated set of experiences with familiar others. Second, people tend to have more incentives for making distinctions among familiar others. Third, people possess a larger sample of exemplars for familiar others and, according to basic statistical principles, greater variability is found among larger samples than among smaller samples. Evidence for the differential intergroup familiarity hypothesis is mixed. Some studies have found supportive evidence (Islam & Hewstone, 1993b; Linville et al., 1989, Experiments 1, 2, & 4), but other studies have found null effects (Askevis-Leherpeux & Bastounis, 1998; E.E.Jones, Wood, & Quattrone, 1981; Park, Ryan, & Judd, 1992) or opposite effects (Oakes, Haslam, Morrison, & Grace, 1995; Ryan & Bogart, 1997). One simple method of testing whether differential familiarity is necessary for OH to occur is to establish whether OH occurs in intergroup situations in which there is no differential familiarity. Minimal groups and gender groups provide two examples of such intergroup situations. Minimal Group Evidence. In the minimal group paradigm (e.g., Tajfel, Billig, Bundy, & Flament, 1971), participants are randomly and anonymously divided into two groups (e.g., “Group A” and “Group B”), ostensibly on the basis of trivial or arbitrary criteria (e.g., preference for art or the toss of a coin). Importantly, these minimal groups are new and unfamiliar to participants, and group members’ personal identities remain anonymous throughout the experiment. Consequently, participants in the minimal group paradigm are as (un)familiar with the in-group and its members as they are with the out-group and its members.1 Hence, evidence of OH in the minimal group paradigm goes against the idea that differential familiarity is necessary for the occurrence of OH (e.g., Judd & Park, 1988).
GENDER OUTGROUP HOMOGENEITY
155
In 1992, Ostrom and Sedikides reviewed the evidence for OH in the minimal group paradigm and concluded that it was “quite meager” (p. 546). However, since this time, evidence for OH in the minimal group paradigm has grown considerably (Badea & Askevis-Leherpeux, 2002; Brewer & Weber, 1994, Experiment 1; Brown & Ross, 1977, Experiment 3, as cited in Ostrom & Sedikides, 1992; Ellemers, Spears, & Doosje, 1997, Experiment 1; Judd & Park, 1988; Lorenzi-Cioldi, 1995b, as cited in LorenziCioldi, 1998; Lorenzi-Cioldi, 1997, as cited in Lorenzi-Cioldi, 1998; Lorenzi-Cioldi, Deaux, & Dafflon, 1998, Experiments 2 & 3; Mackie, Sherman, & Worth, 1993, Experiment 2; Rubin, Hewstone, & Voci, 2001, Experiment 1; J.W.Sherman, Klein, Laskey, & Wyer, 1998, Experiment 1; Simon & Brown, 1987; Simon & Hamilton, 1994, Experiments 1 & 2; Simon & Mummendey, 1990; Simon & Pettigrew, 1990; Wilder, 1980, as cited in Wilder, 1981; Wilder, 1984, Experiment 1; Wilder & Thompson, 1980; Willis, Rubin, & Hewstone, 2000, Experiments 1 & 2). Relatively few minimal group studies have reported null OH effects (Brewer, Weber, & Carini, 1995, Experiment; Brown & Ross, 1977, Experiment 1, as cited in Ostrom & Sedikides, 1992; Doosje, Ellemers, & Spears, 1995, Experiment 2; Doosje, Spears, & Koomen, 1995, Experiments 1 & 2; Ellemers et al., 1997, Experiment 2; Karasawa & Brewer, 1989, as cited in Ostrom & Sedikides, 1992; Simon, 1990, as cited in Ostrom & Sedikides, 1992). Gender Group Evidence. The differential intergroup familiarity hypothesis also predicts little or no OH for gender groups because in most cases men and women have an approximately equal amount of familiarity with one another. Contrary to this hypothesis, there is now a large body of evidence demonstrating gender OH (Cabecinhas & Amâncio, 1999, Experiment 2; Carpenter & Ostrom, 1985, as cited in Messick & Mackie, 1989; Fiedler, Semin, & Finkenauer, 1993; Frable & Bem, 1985; Lorenzi-Cioldi, 1988; Lorenzi-Cioldi, 1992, as cited in Lorenzi-Cioldi, 1998; Lorenzi-Cioldi, 1993; Lorenzi-Cioldi et al., 1995; Mackie, Sherman, & Worth, 1993, Experiment 1; Ostrom, Carpenter, Sedikides, & Li, 1993, Experiment 1; Park & Judd, 1990, Experiment 1; Park & Rothbart, 1982, Experiments 1 & 2; Powlishta, 1995; Young, van Knippenberg, Ellemers, & de Vries, 1999).2 Relatively few experiments have reported null gender OH effects (Beauvais & Spence, 1987; Jackson & Hymes, 1985; Linville et al., 1989, Experiment 3; Linville, Fischer, & Yoon, 1996; Ostrom et al., 1993, Experiment 2; Taylor, Fiske, Etcoff, & Ruderman, 1978, Experiments 2 & 3). In light of the minimal group and gender group evidence cited above, we can conclude that although differential intergroup familiarity may be sufficient to cause OH it is not always necessary (Linville et al., 1989). Given this limitation, the differential intergroup familiarity hypothesis cannot be regarded as providing a convincing explanation of null gender OH effects (e.g., Linville et al., 1989, Experiment 3; Linville et al., 1996). Gender Differences in Gender OH Lorenzi-Cioldi et al. (1995) proposed that (1) overall gender OH effects are unlikely because (2) there are gender differences in gender OH such that men show OH and women show no effect or IH. Contrary to Lorenzi-Cioldi et al.’s first prediction, several studies have found overall gender OH to be a relatively reliable phenomenon (Fiedler et al., 1993; Frable & Bem, 1985; Lorenzi-Cioldi, 1993; Mackie et al., 1993, Experiment 1; Ostrom et al., 1993, Experiment 1; Park & Judd, 1990, Experiment 1; Park & Rothbart, 1982, Experiments 1 & 2; Powlishta, 1995; Young, van Knippenberg, Ellemers, & de Vries, 1999). Nonetheless, there is also evidence consistent with Lorenzi-Cioldi et al.’s second prediction that men show OH and women show no effect or IH. In particular, several studies using the name-matching measure of OH have found that men show significantly greater gender OH than women (Cabecinhas & Amâncio, 1999, Experiment 2; Lorenzi-Cioldi, 1988; Lorenzi-Cioldi, 1992, as cited in Lorenzi-Cioldi, 1998; Lorenzi-Cioldi,
156
THE PSYCHOLOGY OF GROUP PERCEPTION
1993; Lorenzi-Cioldi et al., 1995; Young et al., 1999; see also Frable & Bem, 1985, footnote 4). Furthermore, studies using dispersion measures of OH have found the same pattern of results to be either significant (R. Brown & Smith, 1989) or marginally significant (Linville et al., 1989, Experiment 3; Park & Judd, 1990, Experiment 1). Lorenzi-Cioldi (1998) explains these gender differences in gender OH in terms of differences in group status. Gender and Group Status. People often perceive men to have a higher status than women (Eagly, 1987; Eagly & Wood, 1982; Fiske, 1993; Geis, 1993; Ridgeway & Diekema, 1992; Stewart & Vassar, 2000; Stewart, Vassar, Sanchez, & David, 2000, Experiment 3; Wagner & Berger, 1997). Lorenzi-Cioldi (1998) put forward a variety of social psychological processes to explain how this perceived status difference might account for gender differences in gender OH. We elaborate on three of these processes here. First, Lorenzi-Cioldi (1998, p. 39) proposed that members of high status groups are motivated to enhance the salience of their personal identity in order to claim personal credit for the high status that their group membership entails. This “personalizing strategy” is thought to be a way for group members to legitimize their high status because it implies that their status derives from their personal characteristics rather than from the privileges that their group membership bestows upon them (see also Branscombe, 1998). Hence, members of high status groups should perceive themselves in terms of their personal identity more than members of low status groups do. Some theorists have suggested that the presence of the individuated (personal) self in the ingroup increases perceived ingroup variability relative to perceived outgroup variability (Park & Judd, 1990; Simon, 1993). Hence, greater personal identity salience for members of high status groups may lead them to show greater OH than members of low status groups (Voci, 2000). Second, Lorenzi-Cioldi (1998, pp. 35–36) proposed that people tend to treat members of low status groups in terms of their group membership more than they do members of high status groups. This differential treatment can lead to differences in social identity salience for high and low status group members. Specifically, social identity salience should be higher for members of low status groups than for members of high status groups. As Voci (2000) noted, increasing social identity salience can lead either to perceived equivalence between ingroup and outgroup variability (see also Mullen & Hu, 1989, p. 235) or, under conditions of ingroup threat, to IH (Simon, 1992b). In either case, enhanced social identity salience for members of low status groups should lead them to show less OH than members of high status groups. Finally, Lorenzi-Cioldi and colleagues (1993, 1998; Lorenzi-Cioldi et al., 1995) proposed that high status groups are generally perceived to be more variable than low status groups. This “low status group homogeneity effect” (see Lorenzi-Cioldi, 1998, p. 34; Stewart et al., 2000) is thought to occur independent of the perceiver’s own group membership. In other words, this effect is predicted to be equally applicable for group members and nonmembers alike. The “low status group homogeneity effect” may occur because people in Western societies equate individuality with high status (Snyder & Fromkin, 1980), and it may account for OH because people are motivated to perceive their ingroup as having a higher status than other groups (Tajfel & Turner, 1979). A Matter of Interpretation. Lorenzi-Cioldi and colleagues (Lorenzi-Cioldi, 1998; Lorenzi-Cioldi et al., 1995) predicted an interaction between participants’ gender (male/female) and target group (ingroup/ outgroup) such that men show OH and women show no effect or IH. There are four theoretical interpretations for this interaction, and it is worthwhile considering each in turn in order to appreciate the various confounds that are involved in this type of research. First, it is possible that the interaction between participants’ gender (male/ female) and target group (ingroup/outgroup) reflects a general tendency for men to show greater OH than women, regardless of the target groups involved (Lorenzi-Cioldi et al., 1995; Young et al., 1999). According to this explanation, not only would men be expected to show greater OH than women when the target groups were gender groups,
GENDER OUTGROUP HOMOGENEITY
157
but also when the target groups were minimal groups, occupational groups, national groups, and so on. Relatively few nongender group OH studies have tested for this general gender difference. However, the evidence from these few studies does not support the existence of a general gender difference in OH (Hewstone et al., 1993, Experiments 1 & 2; Simon & Pettigrew, 1990; Wilder, 1990, Experiments 1 & 2; Young, van Knippenberg, Ellemers, & de Vries, 1997). Second, it is possible that the interaction reflects a specific tendency for men to show greater gender OH than women. In other words, men may perceive their gender ingroup to be more variable than their gender outgroup to a greater extent than women. Following Lorenzi-Cioldi (1998), this tendency may be due to men’s enhanced personal identity salience and/or women’s enhanced social identity salience. Third, it is possible that the interaction reflects a general tendency for people to perceive differences in intragender variability: People may tend to perceive men to be more variable than women. Following Lorenzi-Cioldi (1998), this general tendency may represent a “low status group homogeneity effect.” Finally, it is possible that the interaction reflects real differences in intragender variability: Men may actually be more variable than women on the given dimensions (Brauer, 2001). For example, Brown and Smith (1989) suggested this possibility when discussing their findings that male academic staff showed OH and female academic staff showed IH on dimensions related to academic ability. They proposed that sexist employment procedures may have led to standards of selection being higher for female staff than for male staff. These differential selection standards may have led to less actual variability in ability among female staff compared to male staff. Hence, men perceiving OH and women perceiving IH could be explained as accurate perceptions of objective reality rather than as social psychological biases (for similar interpretations, see Brown & Wootton-Millward, 1993; Guinote, Judd, & Brauer, 2002; Simon, GlässnerBayerl, & Stratenwerth, 1991). As Brauer and Judd (2000; see also Bartsch & Judd, 1993; Brauer, 2001) pointed out, the last three interpretations are confounded because the interaction between participants’ gender (male/female) and target group (ingroup/ outgroup) is statistically identical to the main effect of target group when target group is defined as “men/women” rather than as “ingroup/outgroup”. Although measures of actual variability may be used to rule out the last interpretation, the second and third interpretations are particularly difficult to disentangle because no gender-neutral control group is available to provide objective (i.e., unaffiliated) judgments of male and female variability. Since participants are themselves always either male or female, they always have the potential to categorize male and female target groups subjectively (i.e., as “ingroup” and “outgroup”) as well as objectively (i.e., as “men” and “women”). Researchers need to consider other ways of determining the extent to which the interaction between participants’ gender and “ingroup/ outgroup” target group represents differences between participants, perceived differences between “men/ women” target groups, or a combination of both (see Brauer & Judd, 2000). The Effects of Group Status on OH. Lorenzi-Cioldi’s (1998) predictions are applicable to all groups that exist within a status hierarchy, not just gender groups. Consequently, it is instructive to consider the effects of group status on OH in order to evaluate these predictions. Lorenzi-Cioldi (1998, p. 42) acknowledged that OH often occurs between groups of equal status, and he took this “default” OH effect into account when he made his empirical predictions concerning the degree of OH shown by unequal status groups. Specifically, he assumed that the processes associated with high status groups are liable to combine with default OH processes to produce a relatively large OH effect for members of high status groups. In contrast, the processes associated with low status groups are liable to counteract the default OH processes to produce no effect or an IH effect for members of low status groups. Interestingly, Simon and Hamilton (1994, p. 704) arrived at the opposite prediction to Lorenzi-Cioldi (1998) based on the principles of social identity theory (Tajfel & Turner, 1979). They argued that people are
158
THE PSYCHOLOGY OF GROUP PERCEPTION
motivated to perceive themselves in terms of a positive social identity and because high status groups provide a more positive social identity than low status groups, members of high status groups should be more motivated to perceive themselves in terms of their social identity than members of low status groups. It follows that members of high status groups should show less OH than members of low status groups.3 The available evidence provides more support for Lorenzi-Cioldi’s (1998) prediction than for Simon and Hamilton’s (1994) prediction. Members of high status groups tend to perceive greater OH than members of low status groups (Brewer & Weber, 1994; Boldry & Kashy, 1999; Cabecinhas & Amâncio, 1999, Experiment 1; Lorenzi-Cioldi, 1995b, as cited in Lorenzi-Cioldi, 1998; Lorenzi-Cioldi et al., 1998, Experiments 1, 2, & 3; Sedikides, 1997, Experiments 1 & 2; van Knippenberg, Blaauw, & Vermunt, 1996, as cited in Young et al., 1999; Voci & Capozza, 1999), although some studies show no difference in OH between members of high and low status groups (Kashy, 1988, as cited in Boldry & Kashy, 1999), and other studies have found the reverse effect (Brauer, 2001; Judd, Park, Ryan, Brauer, & Kraus, 1995; Pickett & Brewer, 2001). In summary, consistent with Lorenzi-Cioldi (1998), the evidence shows that members of high status groups tend to perceive greater OH than members of low status groups. Consequently, the finding that men show greater gender OH than women may be attributable to perceived intergender differences in status. THE EFFECTS OF GROUP SIZE ON OH Group status and group size are often confounded in the real world such that majority groups are perceived to have a higher status than minority groups (Farley, 1982; Sachdev & Bourhis, 1984, 1991). However, the relationship between group size and group status is not always a positive one. In some cases, a minority group may be perceived as an elite group that possesses higher status than a majority group (e.g., white people in apartheid South Africa). Furthermore, there is some evidence that people may automatically assume high status minimal groups to be numerically smaller than low status minimal groups (LorenziCioldi, 1995b, as cited in Lorenzi-Cioldi, 1998; Lorenzi-Cioldi & Dafflon, 1996, as cited in Lorenzi-Cioldi, 1998). Consequently, it is appropriate to look at the effects of group size independent of group status as well as to consider the possibility of an interaction between these two factors. Theorists agree on the prediction that members of majority groups should show greater OH than members of minority groups, who should show IH (Brewer, 1993; Brown & Smith, 1989; Mullen & Hu, 1989; Simon, 1992). Certainly, the evidence supports the prediction that majority group members show greater OH than minority group members (Badea & Askevis-Leherpeux, 2002; Brewer & Weber, 1994, Experiment 1; Brown & Smith, 1989; Guinote, 2001; Hewstone et al., 1993, Experiments 1 & 2; Simon & Brown, 1987; Simon & Hamilton, 1994, Experiments 1 & 2; Simon & Mummendey, 1990; Simon & Pettigrew, 1990) and that minority group members tend to show IH (Brown & Smith, 1989; Guinote, 2001; Simon & Brown, 1987; Simon, Glässner-Bayerl, & Stratenwerth, 1991; Simon & Hamilton, 1994, Experiment 2; Stephan, 1977; Verkuyten & Kwa, 1996; for a meta-analytic review, see Mullen & Hu, 1989). However, there is little consensus about why this group size effect occurs. Simon and colleagues (Simon, 1992b; Simon & Brown, 1987; Simon & Pettigrew, 1990) proposed that minority group membership poses a threat to self-esteem, and that this threat motivates minority group members to enhance the positivity of their social identity. IH is thought to be one means of enhancing social identity because it indicates that the ingroup provides more solidarity and social support than the outgroup. Brewer (1991, 1993) proposed that social identity represents a compromise between two opposing needs: the need for assimilation and the need for differentiation. Majority group members have a greater need for intra-ingroup differentiation than minority group members because majority groups are, by definition, more
GENDER OUTGROUP HOMOGENEITY
159
inclusive. Consequently, majority group members should be motivated to perceive ingroup heterogeneity (i.e., OH) more than minority group members. In contrast, minority group membership should arouse the need for intra-in-group assimilation, leading to IH. Several theorists have proposed that category salience is greater for members of minority groups than for members of majority groups because minority groups are more visible in the social context (Kanter, 1977; W.J.McGuire & McGuire, 1982; Mullen, 1991; Pichevin & Hurtig, 1996; Simon & Hamilton, 1994; Taylor et al., 1978; van Twuyver & van Knippenberg, 1999). Consequently, social identity salience should be higher for minority group members than for majority group members, leading to less OH for minority group members. Finally, some theorists have suggested that majority groups are generally perceived to be more variable than minority groups (Bartsch & Judd, 1993; Judd, Ryan, & Park, 1991; Messick & Mackie, 1989; Park & Rothbart, 1982). Like the “low status group homogeneity effect,” this “minority group homogeneity effect” (see Park & Rothbart, 1982, p. 1052) is thought to occur independent of the perceiver’s own group membership. The “minority group homogeneity effect” may occur because, on purely statistical grounds, greater variability is observed in large samples than in small samples, and the effect may account for OH because people tend to perceive their ingroup as being numerically larger than their outgroup (LorenziCioldi, 1997, as cited in Lorenzi-Cioldi, 1998; Lorenzi-Cioldi, 1995b, as cited in Lorenzi-Cioldi, 1998; Simon & Mummendey, 1990; Willis et al., 2000, Experiment 1). The Interactive Effects of Group Status and Group Size on OH As we have seen, the evidence indicates that members of high status groups and members of majority groups show the greatest OH effects. But how do group status and group size interact with one another to influence OH? As mentioned previously, Simon and Hamilton (1994) proposed that members of high status groups are more motivated to think of themselves in terms of their social identity than are members of low status groups. However, Simon and Hamilton also proposed that this difference in social identity salience is qualified by group size. Work by Ditto and Jemmott (1989) demonstrated that evaluations of characteristics become more polarized when the characteristics are perceived to be scarce. On the basis of this “scarcity principle”, Simon and Hamilton proposed that group status differences in social identity salience are enhanced for minority group members. Specifically, members of the high status minority should show the greatest IH and members of the low status minority should show the greatest OH. Simon and Hamilton (1994, Experiment 2) manipulated group status and group size orthogonally in a minimal groups experiment and measured OH in terms of perceived intragroup similarity. Contrary to their predictions, they found that members of the high status majority group showed OH, members of the high status minority group showed IH, and members of the low status majority and minority groups showed no significant effects. A subsequent minimal group experiment using a name-matching measure of OH failed to replicate these results (Lorenzi-Cioldi & Dafflon, 1996, as cited in Lorenzi-Cioldi, 1998; see also Ellemers &: van Rijswijk, 1997). Consequently, firm conclusions cannot be drawn with respect to interactions between group status and group size. SOME CONTRIBUTIONS OF OUR OWN RESEARCH In our own research, we were interested in determining the roles of differential familiarity, participants’ gender, and group size in determining gender OH on dispersion measures. Previous research has
160
THE PSYCHOLOGY OF GROUP PERCEPTION
demonstrated gender differences in gender OH using name-matching measures of OH (e.g., Cabecinhas & Amâncio, 1999, Experiment 2; Lorenzi-Cioldi, 1993; Lorenzi-Cioldi et al., 1995; Young et al., 1999). However, to our knowledge, only R.Brown and Smith (1989) found clear evidence for this gender difference using a dispersion measure of OH (the “range” measure). One problem with Brown and Smith’s study is that gender and group size are confounded in its design. Brown and Smith found that majority males perceived greater OH than minority females. This result may be interpreted as evidence for either a gender difference in gender OH (men show greater OH than women) or as a group size difference in OH (majority group members show greater OH than minority group members). In our own research, we used the range measure to investigate gender OH, but we also unconfounded gender and group size in order to analyze each variable separately as well as in interaction with one another (Hewstone & Crisp, 2000; Hewstone, Crisp, Contarello, Conway, Voci, Marletta, et al., 2000; Hewstone, Crisp, Richards, Voci, & Rubin, 2000; Rubin et al., 2001; Rubin, Richards, & Hewstone, 2000; see also Hewstone, Martin, HammerHewstone, Crisp, & Voci, 2001). The first part of our research program provided a competitive test of the gender and group size explanations of Brown and Smith’s (1989) findings. First, we attempted to replicate Brown and Smith’s findings using the same gender-group size relationship that they used. In this experiment (Hewstone & Crisp, 2000, Experiment 2), we asked male and female police officers to judge each group’s dispersion using the range measure. Although we found no overall OH effect (i.e., a main effect of target group), we found a two-way interaction between participants’ gender (male/female) and target group (ingroup/ outgroup). Consistent with Brown and Smith’s results, majority group members (male participants) showed OH and minority group members (female participants) showed IH. Hence, we replicated Brown and Smith’s findings using male majority and female minority groups. Measures of familiarity revealed that majority group members were more familiar with the ingroup than the outgroup and that the reverse was true for minority group members. This pattern of findings makes sense because, all other things being equal, people have a higher probability of interacting with majority group members than with minority group members. Importantly, however, although the familiarity results followed the same pattern as the OH results, familiarity did not mediate OH. Like Brown and Smith’s (1989) study, our first experiment confounded group size and gender in its design. Consequently, the effects that we found could have been due to either gender or group size. In order to disentangle these variables, we conducted a second experiment (Hewstone & Crisp, 2000, Experiment 1) using a population in which women comprised the majority and men comprised the minority. Specifically, we sampled student nurses instead of police officers in order to reverse the gender-group size relationship. Again, although we found no overall OH effect, we found a two-way interaction between participants’ gender and target group. Majority group members (female participants) showed OH and minority group members (male participants) showed IH. Hence, in this second experiment, we replicated Brown and Smith’s group size finding using female majority and male minority groups. Once again, although the familiarity results followed the same pattern as the OH results, familiarity did not mediate OH. Our first two experiments (Hewstone & Crisp, 2000) showed that Brown and Smith’s (1989) findings were most likely to be due to the influence of group size rather than group gender. However, we were also interested in investigating how these two factors interacted with one another. Consequently, we continued our program of research by unconfounding group gender and group size within the same experimental design. In the first of these experiments (Hewstone, Crisp, Richards, et al., 2000, Experiment 1), we sampled male and female university students who belonged to departments that contained either a majority of male students (earth sciences, engineering), a majority of female students (education, psychology), or an
GENDER OUTGROUP HOMOGENEITY
161
approximately equal number of male and female students (biochemistry, mathematics). Participants judged the dispersion of male and female students in their own department. We found a three-way interaction between department (male majority/female majority/parity), participants’ gender (male/female), and target group (ingroup/outgroup). In the male majority and parity departments male participants showed OH and female participants showed no significant effect. In the female-majority departments participants showed no significant effects. Hence, we confirmed Lorenzi-Cioldi’s (1998) prediction of a gender difference in gender OH, but only when participants belonged to male majority or parity departments. Although we found similar results in an additional experiment (Hewstone, Crisp, Richards, et al., 2000, Experiment 3), we failed to replicate the interaction between participants’ gender and group size in a third experiment (Rubin et al., 2001, Experiment 3). In this latter experiment, the typical gender difference in gender OH was not qualified by group size. Contrary to previous research (Brown & Smith, 1989; Linville et al., 1989, Experiment 3), participants in our first experiment (Hewstone, Crisp, Richards, et al., 2000, Experiment 1) rated themselves as being more familiar with their gender ingroup than with their gender outgroup. Again however, we found no evidence of a significant relationship between familiarity and OH. In the experiments reported above (Hewstone, Crisp, Richards, et al., 2000, Experiments 1 & 3; Rubin et al., 2001, Experiment 3), we collected data from students individually. In a further experiment (Hewstone, Crisp, Richards, et al., 2000, Experiment 2), we reasoned that the group size variable would have a greater impact if we collected data from participants in lecture theaters in which members from both gender groups were present. We thought that this experimental setting would enhance the salience of the group size variable leading to stronger group size effects. The results for this experiment were consistent with our predictions. As in our previous experiments, male participants in male majority departments showed an OH effect. However, we also found that female participants in male majority departments showed a significant IH effect. Hence, increasing the salience of the group size variable led to IH for minority female participants (cf. Hewstone & Crisp, 2000, Experiments 1 & 2). Once again, differential familiarity was unrelated to OH. For the sake of completeness, we should also mention two experiments in which we simply asked students to rate the dispersion of male and female university students in general, as opposed to male and female students in their department (Rubin et al., 2001, Experiment 2; Rubin et al., 2000). Although we found overall OH effects, we failed to find a gender difference in either of these experiments. We shall discuss a possible explanation for these null findings a little latter. The last stage of our research program investigated the effects of group size in more detail. Hewstone, Crisp, Contarello, et al. (2000, Experiment 1) sampled male and female staff members at a British university who belonged to departments that contained either a skewed majority of male staff (92% male) or a tilted majority of male staff (73% male; Kanter, 1977). Participants judged the dispersion of other male and female staff members in their department. We found a three-way interaction between department (maletilted/male-skewed), participants’ gender (male/female), and target group (ingroup/outgroup). Men showed OH and women showed IH, and these effects were more pronounced in male-skewed departments than in maletilted departments. In other words, the gender difference in gender OH increased as a function of group size. Hewstone, Crisp, Contarello, et al. (2000, Experiment 2) extended this research by sampling male and female staff members at an Italian university. Some participants belonged to departments that contained either a skewed majority of male staff (92% male) or a tilted majority of male staff (80% male) as in the first experiment. Other participants belonged to a department that contained either a relatively balanced proportion of male staff (60% male) or a tilted majority of female staff (31% male). Again, we found a three-way interaction between department, participants’ gender, and target group. In the male-skewed
162
THE PSYCHOLOGY OF GROUP PERCEPTION
departments, men showed OH and women showed IH. No significant effects were found in the other three types of department. However, we were able to show that perceived variability for men and women varied as a linear function of group proportions: The OH shown by men tended to disappear as they lost their ingroup majority status, and the IH shown by women tended to disappear as they lost their ingroup minority status. Hence, these final two experiments showed that gender differences in gender OH are most likely to occur under skewed group size conditions. To summarize our research findings, we replicated Brown and Smith’s (1989) finding that male majorities show OH and female minorities show IH on the dispersion measure (Hewstone & Crisp, 2000, Experiment 2). However, we expanded on Brown and Smith’s research by showing that this result was primarily due to the group size factor rather than the gender factor (Hewstone & Crisp, 2000, Experiment 1). In a second series of experiments, we unconfounded gender and group size in order to analyze these factors in interaction with one another. We found that gender differences in gender OH only occurred in male majority contexts. Specifically, only men in male majorities showed OH (Hewstone, Crisp, Richards, et al., 2000, Experiments 1, 2, & 3). Women in male majorities were more likely to show IH, especially when group size was particularly salient (Hewstone, Crisp, Richards, et al., 2000, Expt. 2) or skewed (Hewstone, Crisp, Contarello, et al., 2000, Experiments 1 & 2). Hence, we qualified Lorenzi-Cioldi et al.’s (1995) predictions by showing that group size moderated gender differences in gender OH. Taken together, our research findings suggest that group size is a more powerful moderator of gender OH than participants’ gender, and that gender differences in gender OH are most likely to occur when men are in the majority and women are in the minority. CONCLUSIONS Explaining Null Gender OH Effects OH has been found in a wide variety of natural groups. However, there has been some debate over the existence of OH in gender groups (Linville, 1998; Linville et al., 1989; Lorenzi-Cioldi et al., 1995; Park & Judd, 1990). Linville et al. (1989) proposed that gender OH is unlikely to occur because people are equally familiar with members of both gender groups (Linville et al., 1989). However, the extant evidence shows that differential intergroup familiarity is not necessary for OH to occur. Moreover, our own research suggests that even when differences in intergender familiarity do occur, they are unrelated to gender OH effects (Hewstone & Crisp, 2000, Experiments 1 & 2; Hewstone, Crisp, Richards, et al., 2000, Experiments 1 & 2). Consequently, the differential intergroup familiarity hypothesis does not provide a convincing explanation for null gender OH effects. Lorenzi-Cioldi et al. (1995) proposed that overall gender OH effects are unlikely to occur because OH and IH effects shown by men and women counteract one another in the overall analysis. Contrary to Lorenzi-Cioldi et al., there is evidence for both overall gender OH and gender differences in gender OH. Hence, although the proposed gender difference in gender OH appears to be a reliable phenomenon, it cannot be used to explain null gender OH effects. In summary, neither Linville et al. (1989) nor Lorenzi-Cioldi et al. (1995) provide satisfactory theoretical explanations for null gender OH effects. Indeed, one may question whether theoretical explanations are even necessary. Given that null gender OH effects only occur in a relatively small number of studies, and given that the OH effect tends to be a small effect (Mullen & Hu, 1989; Park & Judd, 1990), it is perhaps most parsimonious to explain null gender OH effects as Type II statistical errors.
GENDER OUTGROUP HOMOGENEITY
163
Moderators of the Gender Difference in Gender OH Our research supports Lorenzi-Cioldi et al.’s (1995) prediction of a gender difference in gender OH. However, our findings suggest that this gender differ ence is moderated by three factors: group size, type of dependent measure, and target group description. First, we found that gender differences in gender OH tended to be moderated by group size. Specifically, we found that men only tended to show OH when their gender was in the majority (Hewstone, Crisp, Contarello, et al., 2000, Expts. 1 & 2; Hewstone, Crisp, Richards, et al., 2000, Experiments 1, 2, & 3) and women only tended to show IH when their gender was in a salient minority (Hewstone, Crisp, Contarello et al., 2000; Hewstone, Crisp, Richards et al., 2000, Experiment 2). We believe that membership in male majority groups enhances personal identity salience for men and social identity salience for women (Voci, 2000). Second, with respect to type of dependent measure, gender differences in gender OH tend to have been found using name-matching measures (Cabecinhas & Amâncio, 1999, Experiment 2; Lorenzi-Cioldi, 1988; Lorenzi-Cioldi, 1992, as cited in Lorenzi-Cioldi, 1998; Lorenzi-Cioldi, 1993; Lorenzi-Cioldi et al., 1995; Young et al., 1999; see also Frable & Bem, 1985, footnote 4). However, together with Brown and Smith (1989), our work has shown that this gender difference is equally apparent using the range measure of dispersion. Interestingly, studies using measures of intragroup similarity (Mackie et al., 1993, Experiment 1) or stereotypicality (Park & Judd, 1990, Experiment 1; Park & Rothbart, 1982, Experiments 1 & 2) have failed to find gender differences in gender OH. The obvious factor that links the name-matching and dispersion measures and that separates them from the similarity and stereotypicality measures is that the former two measures require participants to respond in terms of single group members whereas the latter two measures require participants to respond in terms of the group conceived as a whole. We would argue that measures specifying single group members enhance personal identity salience (Voci, 2000), and that men accept this enhancement more than women because they wish to legitimize their high status (Lorenzi-Cioldi, 1998). Finally, the way in which target groups are described also seems to be a potentially important moderator. Consistent with previous research (Brown & Smith, 1989), we found that men showed greater gender OH than women when targets were described as individuals whom participants could relate to on a personal basis (e.g., “male and female students in your department”; Hewstone, Crisp, Richards, et al., 2000, Experiments 1, 2, & 3; Rubin et al., 2001, Experiment 3). However, this gender difference disappeared when targets were described as more abstract social categories that were unlikely to elicit interpersonal comparisons (e.g., “male and female students”; Rubin et al., 2001, Experiment 2; Rubin et al., 2000). Again, we would argue that the potential for interpersonal comparisons enhances personal identity salience, and that men are more prone to this enhancement than women (Lorenzi-Cioldi, 1998). It is interesting to note that some previous research has investigated the influence of target group descriptions on OH. Park and Rothbart (1982, Experiment 1) manipulated target group descriptions (“women [men] as a whole” versus “women [men] in your class”) but found no significant difference between these conditions using a stereotypicality measure of OH. Lorenzi-Cioldi et al. (1995) described male and female targets as either students at the participants’ own university (superordinate ingroup), employees (superordinate outgroup), or persons (control condition). They found the typical gender difference in gender OH whereby male participants showed OH and female participants showed no significant effect. However, this gender difference was qualified by condition such that it was greatest when targets were described as students at the participants’ own university and nonsignificant when targets were described as persons. Hence, consistent with our own results, the gender difference was accentuated when target groups were described in ways that triggered interpersonal as well as intergroup comparisons.
164
THE PSYCHOLOGY OF GROUP PERCEPTION
Summary Overall gender OH is a relatively reliable phenomenon. However, there are gender differences such that men tend to show OH and women tend to show no effect or IH. Although we cannot rule out the possibility of a “low status group homogeneity effect,” we believe that this gender difference in gender OH may arise because men have a higher status than women, and members of high status groups tend to have higher personal identity salience and/or lower social identity salience than members of low status groups. Given a certain degree of social identity salience, enhanced personal identity salience should lead to greater OH because the presence of the personal self in the ingroup increases perceived ingroup variability (Park & Judd, 1990; Simon, 1993). In contrast, enhanced social identity salience should lead either to no difference in perceived ingroup/outgroup variability or, under conditions of ingroup threat, to IH (Mullen & Hu, 1989; Simon, 1992b; Voci, 2000). Notably, the gender difference in gender OH is more pronounced under conditions that enhance personal identity salience, namely (1) membership in a majority group, (2) measures that make reference to single group members, and (3) target group descriptions that encourage interpersonal comparisons. We believe that the enhancing personal identity may exacerbate the gender difference in gender OH because men are more motivated to personalize their ingroup’s status. A MODEL OF OH We conclude by outlining a tentative model of OH and discussing its theoretical implications. Our model of OH assumes that both personal and social identity must be salient in order for OH to occur. Social identity must be salient because perceptions of “outgroup homogeneity” require that people categorize them selves as ingroup members and categorize others as outgroup members. Personal identity must be salient in order to create greater variability in the ingroup than the outgroup (Park &: Judd, 1990; Simon, 1993). It is important to note that, consistent with traditional theorizing (e.g., Tajfel & Turner, 1979; Turner, 1987a), our model assumes that personal identity and social identity cannot be salient at the same time. However, we note three points that support the possibility of an interactive relationship between personal and social identity across time. First, situational factors may have independent influences on personal and social identity salience such that the potential for the salience of each type of identity varies orthogonally from situation to situation. Second, people may be able to switch relatively rapidly between personal identity and social identity within the same situation (Spears, Doosje, & Ellemers, 1999, p. 78). Third, both personal and social identity concerns may influence intergroup perception and behavior (R.Brown & Turner, 1981, p. 65). Taking these three points together, we believe that it is appropriate to consider personal and social identity as having independent and interactive influences on intergroup perception (and behavior) in any given situation. In particular, we predict that OH will occur when situational factors increase the potential salience of both personal and social identity. Our model also attempts to resolve the apparent contradiction between Lorenzi-Cioldi’s (1998) predictions and those based on social identity theory (e.g., Simon & Hamilton, 1994). Consistent with Lorenzi-Cioldi (1998), we predict that members of high status groups will use a personalizing strategy and show greater OH than members of low status groups. However, we also predict that a second set of strategies comes into play when group members make judgements of group dispersion along dimensions that are directly related to the intergroup status hierarchy. Specifically, members of high status groups will be tempted towards IH as a strategic means of consolidating their in-group’s position along the status dimension. In contrast, members of low status groups will perceive greater in-group variability in order to distance their personal self from the in-group’s relatively negative position (Doosje, Spears et al., 1995).
GENDER OUTGROUP HOMOGENEITY
165
This personal dissociation strategy should lead to OH (Rubin et al., 2001). Hence, our model assumes that members of high status groups show OH on status-unrelated dimensions as the consequence of a “personalizing strategy”, and that members of low status groups show OH on status-related dimensions as the consequence of a “dissociation strategy”. We believe that an important avenue for future research will be to investigate dimension type as a potential moderator of the effects of group status on OH. AUTHOR NOTE Part of this chapter was prepared while Miles Hewstone was a Fellow at the Center for Advanced Study in the Behavioral Sciences, Stanford. He gratefully acknowledges the financial support provided by the William and Flora Hewlett Foundation. Correspondence concerning this chapter should be addressed to Mark Rubin at the School of Behavioral Sciences, University of Newcastle, Callaghan, NSW 2308, Australia. Tel: +61 (0)2 4921 6706. Fax: +61 (0)2 4921 6980. E-mail:
[email protected] NOTES 1. Linville et al. (1989, p. 185) proposed that even when there is little or no explicit difference in intergroup familiarity there is still likely to be an implicit difference in intergroup familiarity because people always have privileged knowledge of at least one in-group member: the self. However, in order to explain OH, Linville et al. rely on the additional assumption that the self is represented by more exemplars than other individuals. Consequently, we will treat this “self hypothesis” separately from the “differential familiarity with others” hypothesis. 2. Some of these gender studies found an OH effect that is restricted to male participants (e.g., Cabecinhas & Amâncio, 1999, Experiment 2; Lorenzi-Cioldi, 1992, as cited in Lorenzi-Cioldi, 1998; Lorenzi-Cioldi et al., 1995). However, unless it is assumed that for some reason differences in intergroup familiarity are restricted to male participants, even these limited OH effects provide evidence against the necessity of differential intergroup familiarity. Also, note that Brown and Smith’s (1989) study is not cited in this list of evidence because this study confounded gender with group size, and although differential intergroup familiarity is unlikely to have covaried with gender it is possible that it covaried with group size. 3. Simon and Hamilton (1994, p. 704) originally made their predictions with respect to self-stereotyping but later (p. 709) related these predictions to OH.
12 Group Size, Outcome Dependency, and Power Effects on Perceived and Objective Group Variability ANA GUINOTE University of Kent, UK
Social groups seldom occupy a similar position in society in terms of group size, power and status. Among these variables, many scholars consider power—or the extent to which individuals influence others and control their outcomes—the central variable underlying asymmetric intergroup relations. For instance, power differences among social groups have been pointed out to underlie intergroup discrimination, differences in status between groups (e.g., Jones, 1972; Marger, 1985), and majority-minority intergroup conflict (e.g., DeNardo, 1985; Farley, 1982; Tajfel, 1978). The position groups occupy in the social structure affects group members’ social perception. Recent findings point out that high-status and majority group members perceive their group in a more differentiated and complex way than the outgroup (the outgroup homogeneity effect); whereas low-status and minority group members perceive the outgroup in a more differentiated way than the ingroup (for a review see Devos, Comby, & Deschamps, 1996). Despite the importance of power in asymmetric intergroup relations, research on perceived group variability has focused only on the effects of group size (e.g., Simon & Brown, 1987; Simon & Pettigrew, 1990) and status (e.g., Brauer, 2001; Doosje, Ellemers, & Spears, 1995; Lorenzi-Cioldi, 1993; LorenziCioldi, Eagly, & Stewart, 1995; Sedikides, 1997; Simon & Hamilton, 1994). Surprisingly, no studies directly analyzed the effects of outcome dependency and power on perceived group variability. This chapter discusses recent research that aims to fill this gap, outlining the role of perceived control on group perception. I will first focus on current explanations of asymmetric perceptions of group variability before addressing the effects of power and outcome dependency on group perception.
GROUP SIZE, OUTCOME DEPENDENCY, AND POWER
167
PERCEIVED GROUP VARIABILITY BY MAJORITIES AND MINORITIES Asymmetric perceptions of ingroup and outgroup variability were found in natural groups (Brown & Smith, 1989; Guinote, 2001; Simon, Glaessner-Bayerl, & Stratenwerth, 1991) and laboratory groups focusing either on group size (Simon & Brown, 1987; Simon & Pettigrew, 1990) or on status (Doosje et al., 1995; LorenziCioldi et al., 1995). Mullen and Hu (1989) found in a meta-analysis that the smaller the ingroup relative to the outgroup the greater the tendency to perceive more outgroup than ingroup variability. Research carried out with children shows that the impact of the membership in a social majority/minority starts at an early age: by age 7 white children perceive the white ingroup as more variable than the black outgroup; and black children perceive the white outgroup as more variable than the black ingroup. Similarly, boys perceive the ingroup as more variable than the outgroup, whereas girls perceive the outgroup as more variable than the ingroup (Guinote, Mouro, Pereira, & Monteiro, 2002). Most explanations for perceived group variability by majority and minority group members focus on the effects of group size on ingroup perception. Based on social identity theory, Bernd Simon and Rupert Brown (1987) propose that being a minority causes a threat to self-esteem. Minority members counteract this threat by accentuating solidarity and increasing ingroup similarities, particularly for positive ingroup attributes. Bernd Simon and colleagues suggest that another source of accentuation of ingroup similarities derives from the salience of the minority membership: Because the minority membership is more salient than the majority membership, minority members self-categorize themselves as group members instead of as individuals. They perceive the self in terms of social shared attributes and perceive ingroup members as interchangeable (e.g., Simon & Hamilton, 1994). The Optimal Distinctiveness Theory (Brewer, 1991) offers a third explanation for perceived group variability as a function of group size. This theory suggests that people have two opposite needs whose satisfaction is affected by the size of the ingroup: assimilation with others and differentiation from others. Large groups do not provide the perceiver with their need for differentiation. They are inclined to build subgroups, increasing perceived variability of the ingroup. In contrast, small groups are more likely to fill both needs of assimilation and differentiation, identifying themselves more with their ingroup and accentuating ingroup similarities. Along these lines, Marylinn Brewer (1993) pointed out that to understand perceptions of relative ingroup and outgroup variability we have to direct our attention away from the phenomenon of perceived outgroup homogeneity per se and question the determinants of perceived ingroup homogeneity. These explanations of perceived group variability by majority and minority individuals argue that changes in the perception of relative ingroup and outgroup variability derive from changes in perceived ingroup variability, whereas perceived outgroup variability remains constant. Outgroup members evoke group level perceptions and therefore outgroups are always perceived as homogeneous entities (Haslam, Oakes, Turner, & McGarty, 1995, 1996). Moreover, some authors argue that outgroups evoke distrust and are likely to be stereotyped (e.g., Insko & Schopler, 1998). Without denying that threat, numerical salience, and the need for differentiation and assimilation affect ingroup perception, this chapter points out further explanations that account for asymmetries in perceived group variability. These explanations focus on the effects of perceived control on information seeking and group perception, and on objective group variability. In fact, previous studies focusing on the effects of group size and status on ingroup and outgroup perception confound effects due to the position of the perceiver with effects driven by the target (Barsch & Judd, 1993; Brauer, 2001; Brauer & Judd, 2000). For example, the fact that high-status individuals perceive the ingroup as more variable than the outgroup, and low status individuals perceive the outgroup as more variable than the ingroup is equivalent to saying that
168
THE PSYCHOLOGY OF GROUP PERCEPTION
individuals in both groups perceive the high-status group as more variable than the low status group. This in turn might derive from differences in objective group variability. In this chapter several studies by my colleagues and I will be reported. We conducted these studies to understand the effects of power on group perception, and set apart perceiver and target effects. Overall I will present research addressed to test two major hypotheses: Perceived control affects perceived ingroup and outgroup variability. This hypothesis stems from research on control deprivation, which demonstrates that the extent to which individuals control their outcomes affects attention to social information. Extending these findings to group perception, I hypothesized that perceived control affects perceived group variability, and that the effects of status and group size reported in previous research are mediated by perceived control. More specifically, minorities, low-status and powerless group members lack control over outcomes and increase attention to outgroup information, which in turn increases their perceived outgroup variability (Guinote, 2001). Like the explanations provided above, this explanation also maintains that the position of the perceiver affects group perception, but in this case these effects are extended to outgroup perception and not only to ingroup perception. Furthermore, I will present research supporting the hypothesis that there are differences in objective variability of dominant and subordinated groups, which in turn contribute to asymmetries in perceived group variability. More specifically, perceived control—induced by group size, status, or power—affects the extent to which individuals behave in spontaneous ways. Powerful individuals behave in authentic and individuating ways, whereas powerless individuals are more inhibited and display impoverished action repertoires (e.g., Keltner, Young, Oemig, Heery, & Monarch, 1998; Patterson, 1985; Skinner, 1995). Based on these findings my colleagues and I hypothesized that powerful groups are more variable than powerless groups (Guinote, Judd, & Brauer, 2002). Because perceivers are sensitive to actual group variability (Judd, Ryan, & Park, 1991; Nisbett & Kunda, 1985) they perceive these differences in objective group variability. In other words, one reason why high-power, high-status and majorities are perceived as more variable than low-power, low-status and minorities is because they are objectively more variable. The basis for these explanations and empirical evidence will be discussed in the next sections. First, I will focus on the effects of the position of the perceiver on information seeking and perceived group variability outlining the mediating role of perceived control. Second, I will discuss research supporting the hypothesis that powerful, high-status groups and majorities behave in more variable ways than do powerless, low-status groups and minorities. Furthermore, I will focus on objective group variability within situations (interpersonal variability) and across situations (intrapersonal variability). EFFECTS OF GROUP SIZE, OUTCOME DEPENDENCY, AND POWER ON GROUP PERCEPTION The assumption that lack of control triggers accuracy motivation has its roots in early pragmatic psychology. For instance, William James and John Dewey argued that consciousness and cognitive activity arise when habits do not meet the perceivers’ needs, such as when facing new situations (see Leahey, 1987). A large number of studies support the hypothesis that control deprivation leads to vigilant, effortful information seeking: Low to moderate control-deprived participants actively seek and integrate information (e.g., Weary, Marsh, Gleicher, & Edwards, 1993), engage in explanatory attributions (e.g., Pittman & Pittman, 1980), and are more accurate when facing the attitude attribution paradigm (Pittman, 1993). Motivated perception associated with individual differences also affects information seeking and group
GROUP SIZE, OUTCOME DEPENDENCY, AND POWER
169
perception (Bar-Tal & Guinote, 2002). For instance, lack of control leads low to moderate depressives to display effortful information seeking and relative accurate perceptions (Weary et al., 1993). In the intergroup domain it was found that people often attend to stereotype-consistent information of outgroup members instead of individuated in formation. However, dependency on outgroup members overrides this tendency. When people are dependent on outgroup members they increase attention to individuated information of these members (Erber & Fiske, 1984; Fiske, 2000). This research focused, however, on attention in individual information, and we do not know whether dependency also affects group perception. Walter Stephan (1977) was the first author to suggest that when perceivers need to increase predictability and understanding regarding an outgroup, for instance under threat, they become accuracy motivated regarding the outgroup and increase perceived outgroup variability. Stephan called this phenomenon outgroup differentiation and found support for it in ethnic minority groups. The outgroup differentiation hypothesis, however, has not received attention in research. Furthermore, the factors that lead perceivers to be accuracy motivated regarding outgroups and increase perceived outgroup variability remain unclear. In attempt to fill this gap, I hypothesized that lack of control is a central variable that accounts for outgroup differentiation. When perceivers are dependent on an outgroup— for instance, when they occupy a minority, low-status or low-power position—they increase vigilance regarding outgroup information and become more sensitive to actual outgroup variability. On the other hand, when perceivers have control over their outcomes—for instance, when they occupy a majority, highstatus or powerful position—they do not need outgroup members, pay less attention to them, and perceive them in a homogeneous way (Guinote, 2001). Three important ways in which perceived control over outcomes can operate in majority-minority relations will be discussed in this chapter: group size, outcome dependency, and power. Accordingly, relative group size evokes power relations, such that everything else being equal minorities perceive less control over outcomes than majorities. Second, due to the numerical relations, everyday life of minorities normally involves interaction and interdependence with majority targets. Therefore, minorities tend to be dependent on majorities to achieve outcomes. Power is the third source of asymmetric control: power positions are often occupied by majority citizens and not by minority citizens, giving majorities direct control of important outcomes for minorities. For instance, majority citizens decide whether an emigrant can be given the status of citizenship, whether an ethnic minority member will be accepted to college, whether she or he will receive good grades, and so on. In sum, group size, asymmetric outcome dependency, and power trigger asymmetric control, such that minority, dependent, and powerless group members lack control over their outcomes, whereas majority members, independent group members and powerful group members have control over their outcomes and may directly control others’ outcomes. I will next focus separately on group size, outcome dependency, and power and their effects on perceived group variability. The Power in Numbers Social scientists argue that there is power in numbers (e.g., Della Porta & Diani, 1999; DeNardo, 1985) and that group size can help achieve group goals (Simon, 1998). James DeNardo (1985) maintains that power in numbers explains the dynamics of political parties and social movements and in particular their attempts to increase size. Along these lines, research in social psychology shows that perceivers are influenced by the strength of numbers (Corneille, Yzerbyt, Rogier, & Buidin, 2001; Ng, 1982; Sachdev & Bouhris, 1984,
170
THE PSYCHOLOGY OF GROUP PERCEPTION
1991; Wolf & Latané, 1985; Zdaniuk & Levine, 1996). Therefore, power has often been operationalized in terms of group size (Ebenbach & Keltner, 1998; Ng, 1982). Furthermore, some authors commented that group size affects information processing. For instance, Baron and Misovich (1999) maintain that “minorities, which are smaller and often more involved than majorities, are more likely to engage in systematic thinking” (p. 597). This hypothesis was addressed in a natural minority and majority context (Guinote, 2001). In this study, Portuguese emigrants living in Germany and Portuguese residents living in Portugal described their ingroup and the German out-group in an open-ended format. Answers were analyzed regarding perceived group variability, elaboration, first-hand versus second hand knowledge, and language abstractness. Majority participants perceived more ingroup than outgroup variability and minority participants did the opposite. The source of this interaction relied on outgroup perception, and not ingroup perception. More importantly, minority participants displayed overall more concrete, less trait-related attributes and elaborated more than majority participants. To test the hypothesis that the effects of group size on perceived group variability are mediated by perceived control and elaboration, a further study was conducted (Guinote, Brown, & Fiske, 2000). Participants were categorized into one of two minimal groups: figural and background. Figural group members were described as focusing on the details, being accurate, and analytical. Background group members were described as focusing on the whole, being creative, and flexible. In addition, participants were told that 80% of the population had one style, whereas 20% of the population had the other style. Participants expected to interact with other participants to discuss the relationship between perceptual styles and personal traits. Participants then read exemplar information regarding the four majority and four minority targets they expected to interact with. For each target, two thirds of the information was consistent to the stereotype and one third was inconsistent. Participants were asked to think aloud while reading the information. They then answered a questionnaire with measures of group perception, social identification, and perceived control. As expected, minority members felt less control over their own behavior (they thought that it would be difficult for them to explain their own opinions in the group discussion) and over joint outcomes (they thought that they would influence the direction of the discussion less and that others would listen to them less) than did majority members. An analysis of the think aloud protocols revealed that minority members processed the exemplar information with more effort than did majority members: minority members generated more thoughts, more dispositional attributions, and more explanations than did majority members. In addition, the more participants perceived control the less they elaborated and the more stereotypic were their group perceptions. Perceived control and elaboration completely mediated the effect of group size on the perception of the majority target group—the one who subjectively controls outcomes. The next two studies analyze whether a direct manipulation of control through outcome dependency also affects perceived group variability. Outcome Dependency Leads to Outgroup Differentiation Quite often a person’s outcomes depend on outgroup members. An example of intergroup interdependence is when students in different academic majors work together on a class project. Intergroup outcome dependency is of particular relevance for the understanding of the mechanisms that lead to outgroup differentiation. If short-term task dependency affects group perception under conditions where groups have the same status, size, and power, asymmetries in the groups positions are not necessary to induce
GROUP SIZE, OUTCOME DEPENDENCY, AND POWER
171
asymmetric perceptions of group variability. Instead, the current short-term, control-related needs of perceivers are enough to induce these effects. Several studies show that mere dependency increases attention (e.g., Berscheid, Graziano, Monson, & Dermer, 1976). In particular, dependency on outgroup members increases attention toward them (e.g. Bogart, Ryan, & Stefanov, 1999; Erber & Fiske, 1984; Fiske, 1993; Goodwin, Gubin, Fiske, & Yzerbyt, 1999; Neuberg & Fiske, 1987). Fiske and colleagues argue that this occurs through increased attention to stereotype-inconsistent information, dispositional (personal) inferences, and attribute-based evaluation (as opposed to category-based). Two studies were designed to address the effects of outcome dependency on group perception (Guinote & Fiske, 1999, study 1). In the first study participants were expected to generate ideas for a video about life on campus. For this task they were supposed to collaborate with outgroup members: communication students. In the low-dependent condition participants were supposed to write an individual proposal for the video. In the high-dependent condition participants were expected to write a group proposal with the communication students. Participants then read information about six communication students; each was described with five stereotype consistent attributes and one stereotype inconsistent attribute of the communication students. Finally, participants answered a questionnaire that assessed their perception about communication students in general. As expected, participants whose outcomes depended on the communication students perceived communication students as more variable, less stereotypic, composed of more subgroups than did participants whose outcomes did not depend on the communication students. Importantly, these effects occurred for the perception of communication students in general over and above the six targets participants expected to encounter. Outcome dependency lead, therefore, to increased perceived outgroup variability. A second study extended these findings to include perceptions of ingroup variability and manipulate whether participants were dependent on the ingroup or the outgroup (Guinote & Fiske, 1999, study 2). Following a minimal intergroup procedure, participants were led to believe that they would work on a proposal for an ecological project with either ingroup members or outgroup members. Before working on the proposal participants read information regarding four ingroup and four outgroup targets; each was described with consistent information (2/3) and inconsistent information (1/3). After reading the information participants answered a questionnaire with measures of ingroup and outgroup perception. In this study participants perceived more variability of the target group they were dependent on: ingroup dependent participants perceived the ingroup as more variable than the outgroup, whereas outgroup dependent participants perceived the outgroup as more variable than the ingroup. The next two studies focus on lack of control induced by outgroup members directly controlling outcomes. Power and Perceived Group Variability Power has been defined as potential influence (Lewin, 1941), and as control over outcomes (Thibaut & Kelley, 1959). A great deal of research in this domain concerns whether and under which conditions powerful persons abuse power (Chen, Lee-Chai, & Bargh, 2001; Knippis, 1972). Some researchers argue that powerful people pay less attention to their subordinates. For instance, when subordinates talk powerful individuals look less at them (Dovidio, Ellyson, Keating, & Heltman, 1988) and often interrupt them (Bales, 1953; Georgesen & Harris, 2000). Moreover, powerful individuals attend more to stereotypeconsistent than inconsistent attributes of their subordinates (Goodwin et al., 1999). Powerless people on the other hand pay attention to powerholders (Dovidio et al., 1988). Therefore, it is possible that individuals who are controlled by a group of powerful targets, are motivated to differentiate them from one another to
172
THE PSYCHOLOGY OF GROUP PERCEPTION
increase predictability and control (for a different position see Fiske & Dépret, 1996; Stevens & Fiske, 2000). As a consequence they would increase perceived outgroup variability. Charles Judd, Markus Brauer, and I (Guinote et al., 2002) examined this hypothesis in two studies. These studies included judgments of individual and group perception from participants who were themselves in a powerful or a powerless position and from external observers. Participants in study 1 were told that we were interested in studying how people find creative solutions and how people judge creativity. Participants were randomly assigned to the powerful and the powerless conditions, with four participants in each group. We called the first group the judges and the second group the workers. Workers were expected to work on tasks and judges were expected to evaluate these tasks. Depending on this evaluation, workers would earn between $6 and $10. Judges were paid $8. Judges controlled workers’ outcomes. Judges and workers engaged in a training task and introduced themselves to the group. They were videotaped while working and while introducing themselves. Participants then answered a questionnaire with measures of individual and group perception. In a second study observers saw a videotape with one powerful and one powerless group from study 1. Half of the observers knew about the power relations between the groups. The other half of the observers did not know: they where told that they would see two groups of people randomly selected from the larger pool of participants in our study. Observers answered the same questionnaire as participants from study 1. In the first study powerless individuals perceived the powerful target group as more variable than the powerless target group, whereas powerful individuals displayed symmetric perceptions of ingroup and outgroup variability. In the second study observers rated the powerful group as more variable than the powerless group. Powerless perceivers and observers had, therefore, similar perceptions of the target groups, but powerful perceivers did not. One conclusion that we can derive from this finding is that the position of the perceiver affected group perception, and that powerless individuals were more accurate in their group perceptions than were powerful individuals. Furthermore, powerlessness reversed the outgroup homogeneity effect similar to group size and status. Overall, these findings associated with group size, outcome dependency, and power provide evidence for the position of the perceiver affecting vigilance and group perception. I will focus now on the effects of these variables on objective group variability. THE EFFECTS OF POWER AND GROUP SIZE ON OBJECTIVE GROUP VARIABILITY Previous research provides indirect support for the hypothesis that powerful, high-status and majority group members act in more diverse, and individuating ways than powerless, low-status and minority group members. For instance, findings in animal research indicate that dominant animals initiate more fights, more plays, and more exploratory behavior compared to non-dominant animals (Maslow, 1973a; 1973b). Similarly, children who are not given controlling limits are more spontaneous and make more creative and diverse paintings than children who are given controlling limits (Koestner, Ryan, Bernieri, & Holt, 1984). Furthermore, when given the opportunity to make behavioral choices, non-controlled children make more choices than children who are controlled. The latter children become more prone to allowing others to make future behavioral choices for them than the former (Haddad, 1982). Research focusing on non-verbal behavior points out that powerful individuals act in more authentic and spontaneous ways, reflecting their internal states, than powerless individuals (Hecht & LaFrance, 1998; Patterson, 1985). In addition, powerful individuals display action repertoires that are richer than those of powerless individuals (Skinner, 1995).
GROUP SIZE, OUTCOME DEPENDENCY, AND POWER
173
Scholars of economics came to similar conclusions: focusing on well-being they found that not only do the poor have fewer choices or less advantages than other people but they also underfunction. This means that they succeed less well in doing with the commodities and characteristics at their command (Sen, 1985). In sum, research focusing on animal and human dominance, on lack of control in children, on individuals in chronic power positions, and on larger segments of society suggests that powerless individuals behave in constricted ways, whereas powerful and independent individuals are more disinhibited and behave in more individuating ways. Most of these studies focused on individuals in naturally occurring power positions, instead of studying experimentally the effects of power on behavior. Furthermore, they did not directly analyze the effects of power, status, and group size on objective group variability and its consequences for perceived group variability. My colleagues and I designed studies to address these issues. The two studies by Guinote, Judd, and Brauer (2002) reported above tested the hypothesis that the position of the perceiver in a powerful or a powerless group affects perceived group variability as well as two other hypotheses. One hypothesis maintains that powerful groups are objectively more variable than powerless groups, which in turn contributes to the differences in perceived group variability. The other hypothesis focuses on the effect of knowledge of the targets’ power on group perception (for details regarding this latter effect see Guinote et al., 2002; see also Lorenzi-Cioldi, 1998). In these studies, support for powerful groups being objectively more variable than powerless groups is found if observers who do not know the power of the targets rate the powerful target group as more variable than the powerless target group. In fact, this was what occurred. More specifically, regardless of their knowledge about power of the target groups, observers judged the powerful group to be more variable than the powerless group. This occurred in terms of the range and the standard deviation of the ratings given to individual targets on several attributes. In addition to the traditional measures of perceived group variability, we used the social relations model (Kenny, 1994) to estimate perceived variability from individual judgments (see Boldry & Kashy, 1999). We found that regardless of knowledge about power of the target groups, observers were better able to differentiate one target individual from another in a consensual way (high target variance) and in idiosyncratic ways of perceiving people (high relationship variance) when the target group was powerful than when it was powerless. Higher objective variability of powerful individuals derived from higher behavioral disinhibition and higher individuation: for example, self-presentations of powerful individuals were longer in time, more individuating, and more varied in interests and attributes than powerless individuals’ self-presentations. Overall, the results of these two studies support the hypotheses that: 1) the position of the perceiver affects group perception, with powerless individuals differentiating the outgroup more than the ingroup and having more accurate perceptions of group variability than powerful individuals, and 2) powerful individuals behave in more variable ways than powerless individuals, which in turn contributes to the differences in perceived group variability. The hypothesis that knowledge of power would affect perceived group variability did not receive support. These studies examined objective group variability within one situation, in which group members were working together and presenting themselves. Real life, however, is dynamic. Individuals are exposed to different situations and these situations affect behavioral consistency (Mischel & Shoda, 1995). Therefore, the hypothesis that perceiver position affects objective group variability would be strengthened if we showed that high-power, high-status, and majority group members behaved more variably across situations than do low-power, low-status and minority group members.
174
THE PSYCHOLOGY OF GROUP PERCEPTION
Objective Cross-Situational Variability Previous research demonstrates that situations are powerful determinants of behavior. Gibson (1977; see also Kaplan, 1978) referred to affordances as the potential uses or activities a place suggests to observers by virtue of its physical properties. The impact of situations has been examined in research on personality (e.g., Mischel & Shoda, 1995), aggression (Andersen, 1997), verbal communication (e.g., Giles, Coupland, & Coupland, 1991), non-verbal behavior (e.g., Chartrand & Bargh, 1999), and physical environments’ effects on behavior (e.g., Aarts & Dijksterhuis, 2001). In all these areas, findings reveal that individuals tend to shift their behavior toward the social and physical characteristics of the environment (but see Giles et al., 1991, for instances of divergent verbal accommodation). Fitting into the environment is adaptive. In interaction settings, for example, it enhances communication and liking between partners (Chartrand & Bargh, 1999). Recent findings reveal that situational effects on behavior occur, to a great extent, automatically. For instance, individuals primed with the concept of library automatically talked more quietly than individuals exposed to a neutral prime (Aarts & Dijksterhuis, 2001, study 2). Similarly, individuals who were primed with the name of an exclusive restaurant let less crumbs fall when eating crackers than individuals exposed to a neutral prime (Aarts & Dijksterhuis, 2001, study 4). Studies focusing on mimicry reveal similar results. For example, Tanya Chartrand and John Bargh (1999) led participants to talk to a confederate who either touched her face or rocked her feet. Participants automatically mimicked the confederate’s behavior. These authors called this phenomenon the chameleon effect, after the animal that changes colors according to its physical environment. In spite of this general tendency to fit to environmental cues, several factors can override automatic control of behavior by the environment. In particular, motivation, self-focus, and any shift in attention toward current functioning override automatic responses to the environment (Dijksterhuis, Bargh, & Miedema, 2000). This, in turn, makes individuals less responsive to salient environmental cues that they otherwise would respond to without awareness. For instance, they might stop imitating a partner’s posture or stop behaving in stereotype consistent manners after being exposed to a prime that activated stereotypes (Dijksterhuis, et al., 2000; Dijksterhuis & van Knippenberg, 2000). Yaacov Trope, Cristina Murinho, and I examined whether behavioral variability across situations was moderated by the position of the perceiver in the social structure (Guinote, Trope, & Murinho, 2002). Majority, powerful, and high-status individuals have less constraints, and are therefore in a better position to resist influence than to minority, powerless, and low-status individuals (Lewin, 1941). The former individuals might, therefore, feel free to behave as they wish to and feel less inclined to fit to the concrete situations they encounter than the latter individuals do. This would lead them to display less crosssituational variability than minority, powerful and high-status individuals. However, situational influences do not diminish individual expression. “Researchers in this vein have systematically compared this power of the situation to that of the person as if the two were competing in a zero-sum game, in which evidence for the power of one necessarily diminishes that of the other” (Mischel, 1997, pp. 181). According to interactionist perspectives, individuals can respond highly to a situation and still act in highly individuating ways. Furthermore, we hypothesize that the ways power, status, and group size affect information processing affect in turn responsiveness to situations. Majority, powerful, and high status individuals focus on salient aspects of their target of judgment, whereas minority, powerless, and low status individuals increase vigilance, in particular, to non-salient aspects. This leads the latter to have further cues for action control and respond less to central affordances. When making decisions about their course of action, powerless, lowstatus, and minority individuals might tend to weigh information, be cautious, and respond in less extreme
GROUP SIZE, OUTCOME DEPENDENCY, AND POWER
175
ways to the situation, whereas powerful, high-status, and majority individuals might respond more to the central aspects of the situation and therefore change more from one situation to another. Trope, Murinho, and I designed three studies to test these hypotheses. In the first study participants were assigned to a minimal majority or minority group. Participants were run individually. They were told that the goal of the study was to analyze how perceptual styles affect the interpretation of everyday life situations. After being given feedback about perceptual styles and the sizes in the population that displayed each style (80% vs. 15%) the participant talked with one confederate who was trained to behave in one of two styles: formal and informal. The participant’s task was to tell stories about pictures presented by the confederate. After talking with the first confederate the participant talked with the second confederate. The second confederate was trained to display a different style than the first confederate. In the formal style the confederate was dressed up, acted in a formal way, and was articulate. The confederate in the informal style, was expressive, dynamic, and less articulate. The participant, but not the confederates, was videotaped while talking with the confederates. Later, observers unaware of the participants’ group size rated each individual participant on several traits related to confederate styles. As expected, participants displayed behaviors that corresponded to the style of the confederate they were talking to. When the confederate was formal and well articulate, participants were more formal, more articulate, and less expressive. When the confederate was informal, expressive, and less articulate, participants acted in more casual ways, were more expressive, and less articulate. More importantly, these differences were stronger for participants in the majority group compared to participants in the minority group. In spite of the greater freedom of majority individuals, they were more controlled by the environment and behaved more variably across situations than were minority individuals. This study focused on the domain of interpersonal interaction, which has important automatic components (Chartrand & Bargh, 1999). Therefore, it is also important to examine whether these results are also true when individuals have the opportunity to ponder and make decisions about their course of action. Three studies were conducted with this goal. In one study, participants who were members of a numerical majority or minority imagined themselves being in one of two situations. They were either doing an internship in a firm or had a friend visiting at their house. Participants described how two days in their lives would look from morning to evening while they are in that situation. Participants’ open-ended answers were coded in regard to the extent to which they referred to work-related or social-related activities. In the internship situation work-related activities were those related to the internship. In the friend visiting situation, work-related activities were attending classes, studying, or doing homework. Social-related activities were any activities that participants did for leisure with the visiting friend or with friends in general. As hypothesized, we found that participants in the internship situation reported more work-related activities than social-related activities, whereas participants in the friendship situation reported more socialrelated activities than work-related activities. Furthermore, this difference was stronger for partici pants in the majority condition, compared to participants in the minority condition. This finding indicates that even when individuals have the opportunity to think and make decisions about their behavior, majority individuals report behaving in more variable ways across situations than do minority individuals. Similar results were obtained in a further study manipulating power instead of group size. In another study participants were assigned to a powerful or a powerless group, and were asked to imagine one day in the future and describe this day from morning to evening. In this study we manipulated the situation by having participants imagine one day during the week and one day during the weekend. We analyzed the proportion of leisure and routine activities they reported. As expected participants reported
176
THE PSYCHOLOGY OF GROUP PERCEPTION
more leisure than routine activities during the weekend and more routine than leisure activities during the week. Furthermore, this difference was stronger for powerful than powerless participants. Overall these studies demonstrate that majority and powerful individuals display greater variability in behavior across everyday life situations than do minority and powerless individuals. Although powerful individuals might be efficient in opposing the situation if this is a salient goal for them (driven by personal will), our results indicate that often the affordance of the situation is the most salient cue that guides behavior. Because powerful individuals respond more to salient cues and less to non-salient ones, they respond more to what is central in situations, changing behavior from one situation to another. DISCUSSION This chapter focused on the effects of group size, outcome dependency, and power on perceived and objective group variability. The line of research presented here contributes to the understanding of the outgroup homogeneity effect by majority, high-status, and high-power group members and the reversal of this effect by minority, low-status, and low-power group members. Several studies support the following explanations for asymmetric perceptions of group variability. The effects of group size on group perception are mediated by perceived control. Not surprisingly, a direct manipulation of control through outcome dependency and power held similar results as group size and status. Powerlessness reverses the outgroup homogeneity effect. Interestingly, even when target groups occupy similar positions in the social structure, short-term outcome dependency has similar effects as powerlessness. Up to date the reversal of the outgroup homogeneity effect was only associated with group size and status. Therefore, these findings bring a new perspective into the field. Two mechanisms contribute to asymmetric perceptions of group variability. The first one proposes that the position of the perceiver in a majority, powerful, or high-status group or in a minority, powerless, or lowstatus group affects perceived control, information seeking, and perceived group variability. More specifically, minorities, low power, and low-status group members lack control, are vigilant toward outgroup information and increase perceived outgroup variability. On the other hand, majority, high-power, and high-status group members need members of the outgroup less, and therefore pay less attention to them and perceive more ingroup than outgroup variability. Although support for these hypotheses was found in a wide variety of contexts, which include natural groups and experimental manipulations of group size, outcome dependency, and power, exceptions to these patterns can be found in other contexts. For instance, powerful individuals who feel responsible for subordinates (Overbeck & Park, 2001) or have a communal orientation (Chen et al., 2001) are not likely to stereotype subordinates. Similarly, under certain conditions, minority, powerless, and low-status group members might perceive outgroup members in a stereotypic manner and the outgroup as homogeneous (see Corneille & Yzerbyt, 2002). Previous research suggests that under illegitimacy (Tajfel, 1978) and intergroup conflict or threat (Corneille et al., 2001; Reynolds, Oakes, Haslam, Turner, & Ryan, this volume; Rothgerber, 1997) dependent group members perceive outgroups in a stereotypic manner. Furthermore, minorities might organize themselves and engage in primary strategies to regain control, instead of adopting secondary strategies based on information seeking and accuracy motivation. Activism and social movement participation by minorities—for instance, “Black is Beautiful” and “Gay Pride”— reflect such strategies (Della Porta & Diani, 1999; Simon, 1998; see also Reynolds et al. this volume). These strategies normally occur when group boundaries are not permeable and intergroup relations are perceived as illegitimate. Furthermore, social change requires awareness, plus minorities’ ability to organize themselves, actively oppose the social system, and oppose socially shared beliefs.
GROUP SIZE, OUTCOME DEPENDENCY, AND POWER
177
Unfortunately, in spite of being discriminated against, minorities oftentimes do not have this awareness. Research indicates that minorities tend to have an assimilative tendency. For instance, minority individuals often attribute failure to themselves and are less likely to perceive and act against individual discrimination. For example, more white employees filed race discrimination complaints against the U.S. government in 1996 than did black employees (Harris, 1997). Similarly, at the cognitive level, minorities are more likely to conform to others’ opinions and attitudes than majorities (Asch, 1956). It is not a coincidence that the first 20 years of research devoted to asymmetric social influence focused only on this tendency, and only later did social researchers devote attention to active expressions of minorities and their attempts to change the majority (e.g., Moscovici, 1976). The membership in a dominant or in a subordinate group leads not only to accuracy motivation and outgroup differentiation, but also affects behavior. The second explanation discussed in this chapter to explain asymmetric perceptions of group variability argues that these perceptions are, in part, due to highpower, high-status, and majority group members acting objectively more variably than low-power, lowstatus, and minority group members. This derives from enhanced control leading to increased responsiveness to central aspects of situations and lack of control leading to responsiveness to non-salient aspects of situations. Two studies found support for these differences within a situation (focusing on variability between members of a group). In these studies powerful individuals responded more unequivocally to a self-presentation task. Therefore, their inter-individual differences became salient to observers. Four other studies found support for differences across situations (focusing on intrapersonal variability). These studies indicate that powerful, high-status, and majority group members respond more to the central demands of situations than do powerless, low-status, and minority group members. Therefore, the former individuals change more from one situation to another than the latter. Behavioral variability within and across situations are two different sources of objective variability that contribute to higher perceived variability of powerful, high-status, and majority groups compared to powerless, low-status, and minority groups. Several consequences derive from these findings. Individuals who are control-deprived display high cognitive activity, high vigilance, behavioral inhibition, and poor adaptation to situations, even when situations are not directly related to control deprivation. Because fitting into the demands of the situation is adaptive, minority, low-power, and low-status group members have a deficit at this level. The poor performance of minorities in several areas, including academic performance (e.g., stereotype threat; Steele, 1997) can be understood as a reflection of this general behavioral deficit. The current findings can also contribute to our understanding of minorities low participation in opportunities given to them to enhance their life quality and well-being (e.g., Sen, 1985; 1989). For instance, in spite of affirmative action, whites are more likely to be part of the pool of candidates in the recruitment stage of high-status jobs than are blacks (Braddock & McPartland, 1987; Pettigrew & Martin, 1987). Considering the results presented here, low participation might be in part due to inhibition induced by the position they occupy in the social structure. AUTHOR NOTE Preparation of this chapter was supported by grant POCTI/PSI/43319/2001 from Fundacao para a Ciencia e a Tecnologia. Correspondence should be addressed per email to:
[email protected]. I thank Jason Lawrence, Markus Brauer, Bernadette Park, Yacoov Trope and Alexander Todorov for comments on an earlier draft of this chapter.
13 The Acquisition, Transmission, and Discussion of Social Stereotypes Influences of Communication of Group Perceptions MARKUS BRAUER University of Clermont-Ferrand CHARLES M.JUDD MICAH S.THOMPSON University of Colorado at Boulder
BACKGROUND AND HYPOTHESES Park and Hastie (1987) distinguished two modes of stereotype acquisition. Some stereotypes are primarily “instance-based,” that is, they are acquired through direct contact with individual members of the group. Representations of these individual group members and of the ways they behave are an intrinsic part of the stereotype. As examples, one may imagine a child who grows up in a interracial neighborhood or a businessperson who travels to a country she has never been to before and about whose inhabitants she does not have particular expectations. Other stereotypes are primarily “abstraction-based,” that is, they are communicated to the perceiver by significant socializing agents. These stereotypes consist largely of general propositions about what the group as a whole is like. They are based on no contact (or very limited contact) with members of the target group. For example, one may imagine a European American child who grows up in an area of the country where there are very few minorities but whose parents told him what “blacks” or “Latinos” were like. Alternatively, one may imagine a businessperson who travels to a country to which many of her colleagues have been before and about which she has heard a lot. Although most stereotypes in real life have probably been acquired both through direct contact and communication, there nevertheless exist differences in the extent to which they are primarily instance-based or primarily abstraction-based. Park and Hastie (1987) argued that the mode of stereotype acquisition should influence the perceived variability of the group. Given that abstractionbased stereotypes are most likely communicated as abstract descriptions that describe the group as a whole, the perceived variability within the group should be relatively low. Instance-based stereotypes, however, contain not only group-level information but also
THE ACQUISITION, TRANSMISSION, AND DISCUSSION OF SOCIAL STEREOTYPES
179
representations of individual group members. As a result, perceivers should see the target group as being more variable, with differences among the individual group members. To test their idea that abstraction-based stereotypes contain less variability information, Park and Hastie (1987, Experiment 2) asked participants to form an impression of a hypothetical group that was described as generally intelligent and sociable. Participants received two kinds of information: Abstraction-based information (three abstract descriptions of what the group was like, written supposedly by three individuals who knew the group well) and instance-based information (28 instances of behaviors that specific group members had presumably engaged in). These two kinds of information had been pretested to show that they both conveyed exactly the same mean impression of what the group was like. All participants received both kinds of information, but half of the them first read the abstraction-based information and then the instancebased information, whereas the other half of the participants received the information in the opposite order. Consistent with their hypothesis, Park and Hastie found that participants in the abstraction-first condition perceived the group to be significantly less variable than participants in the instance-first condition. There were no condition differences in the perceived central tendency of the group nor in participants’ memory for confirming or disconfirming behavioral instances, suggesting that the mode of stereotype acquisition affected the perceived variability of the group but not the degree to which the various pieces of information were learned. The work by Park and Hastie (1987) suggests that perceivers who learn about a group through communication may perceive the group to be less variable than if they learn about the group based on firsthand knowledge. But this conclusion remains somewhat speculative, since Park and Hastie did not actually study communication. They constructed the stimulus materials that were used to acquire the group stereotype. The goal of the first two studies we conducted was to examine our first hypothesis according to which similar effects would be found when group abstractions were actually constructed by participants and then communicated from one to another. Downing, Judd, and Brauer (1992) demonstrated that the activity of repeatedly expressing one’s attitude causes the attitude to become more extreme (see also Brauer, Judd, & Gliner, 1995). This effect is particularly strong when others appear to be influenced by the resulting communication (Brauer et al., 1995). Putting one’s attitude into words, talking out loud, defending a point of view, and knowing that others may be influenced by what we say, may cause our attitudes to become more streamlined, less cognitively complex, and more extreme (Judd & Brauer, 1995). Applied to stereotyping, one might argue that the activity of telling other people about what a group is like should influence the extent to which the communicator subsequently sees the group as relatively homogeneous. Other studies provide support for this idea. Allport and Postman (1947), in their work on rumor transmission, argue that speakers tend to transmit information that is less detailed, more concise, and less qualified than the information they have heard (see also Kashima, 2000b). Other researchers have shown that the anticipation of communicating information to others causes individuals to alter the representation of the information they possess (Brock & Fromkin, 1968; Cohen, 1961; Harkins, Harvey, Keithly, & Rich, 1977; Harvey, Harkins, & Kagehiro, 1976; Leventhal, 1962; Zajonc, 1960). Compared to participants who expect to receive further information on some subject, those who anticipate transmitting the information they have already received develop more polarized and simplified representations of that information (E.T.Higgins, 1992; E.T. Higgins, McCann, & Fondacaro, 1982). The work on repeated attitude expression and on anticipated communication provides joint evidence for a second hypothesis that we explored in the studies reported below. Based on this work, we expect that the process of communicating a group stereotype to others should influence the communicator’s stereotype of
180
THE PSYCHOLOGY OF GROUP PERCEPTION
the group. Subsequent to communication, the group should be perceived to be less variable as well as more extremely stereotypic. The above two hypotheses, one concerning the consequences of stereotype acquisition from others and the other concerning consequences of stereotype transmission, address only some components of the communication situation. Both involve a one-way communication, either from a socializing agent to the perceiver or from the perceiver to an audience. In daily life, communication about social stereotypes typically involves an exchange of information and arguments between interacting partners. Along these lines, Moscovici and Zavalloni (1969) asked groups of French students to discuss their attitudes toward Americans. They found that initially unfavorable attitudes were even more unfavorable after the group discussion. This polarization occurred both on judgments reached by the group as a whole and on individual group members’ subsequent judgments. Similar results were obtained by Myers and Bishop (1970) when they formed homogeneous groups of relatively high-, medium-, or low-prejudice participants and asked them to discuss their racial attitudes. As expected, discussion with like-minded others polarized subsequent attitudes, thereby increasing the magnitude of the difference between high- and low-prejudice participants in their evaluative ratings of African Americans. Moscovici and Zavalloni (1969) and Myers and Bishop (1970) examined only one aspect of stereotypes, namely the extremity of the group perception. They showed that the target group, as a whole, was seen more negatively or more positively after group discussion. Given our earlier hypotheses, it also seems likely that informal discussions of what a group is like with like-minded others may also affect the perceived variability of the group. Nearly all theoretical accounts of discussion-induced group polarization (see Brauer & Judd, 1996, for a review) suggest that arguments that are consistent with the predominant attitude in the group are more likely to be mentioned in the discussion, have a greater chance of being considered as valid evidence, and have a greater impact on postdiscussion attitudes than inconsistent arguments. Discussion, therefore, should focus perceivers’ attention on stereotype-consistent information and withingroup similarities rather than stereotype-inconsistent information and differences among group members. Moving beyond one-way communication settings then, it seems likely that informal group discussion, where there is real give-and-take to the communication process, is likely to influence subsequent stereotypes, affecting both the extremity of the group stereotype and the perceived variability of the group members to each other. The third and fourth studies that we report below explore this possibility. Recent work on small groups has shown considerable interest in the effects of information distribution on group judgments. Following Burnstein and Vinokur’s (1977) lead, Stasser and Titus (1985) distinguished between two kinds of prediscussion information distributions: shared information is familiar to all group members, whereas unshared information is held by only one or more (but not all) group members. In the realm of group perception in the real world, information that is consistent with the stereotype is likely to be shared by all members of an informal discussion group. After all, it is hard to avoid being exposed to the cultural beliefs about what, for example, “African Americans” in the United States or “Arabs” in France are like. However, information that is inconsistent with the stereotype is not necessarily shared. In some discussion groups, all members may be familiar with a moderate amount of information that contradicts the cultural stereotype (here the inconsistent information is “shared” across members of the discussion group). In other discussion groups, a minority of members may have had access to numerous instances of behaviors that are inconsistent with the cultural stereotype, whereas other members may not know of such instances (here the inconsistent information is largely “unshared” because only some of the discussion group members are familiar with it). Given that groups are generally quite ineffective in pooling unshared information (Stasser & Titus, 1985; 1987), one may wonder whether the effects of discussion on subsequent stereotypes are moderated by the distribution of information that is inconsistent with the cultural stereotype.
THE ACQUISITION, TRANSMISSION, AND DISCUSSION OF SOCIAL STEREOTYPES
181
On the one hand, if unshared information is not pooled during group discussions, then it might seem that inconsistent information will have little impact on the discussion and subsequent stereotypes about the target group when only some of the discussion group members are familiar with it (i.e., when the stereotype inconsistent information is “unshared”). This would suggest that the effects of group discussion on stereotypes ought to be larger in this unshared condition, since information that might weaken the group stereotype would tend not to be discussed. On the other hand, Vinokur and Burnstein (1978) reported that members of discussion groups have the tendency to put forward arguments that support their prediscussion preferences. When the pieces of counterstereotypic information are concentrated in one group member and are, thus, “unshared,” that one group member is most likely to have a less stereotypic view of the target group. The other group members probably hold a more stereotypic view given that they know less counterstereotypic information. During the discussion, the group member who has the most counterstereotypic information explains why she has a mixed attitude, and the other may argue against it. The motivation to examine the validity of the counterstereotypic information is relatively high. In order to arrive at a common understanding of the target group, members shift their appraisals in the direction of the information discussed. As a result, no polarization (or very little polarization) occurs, and group members converge on a position that is somewhere in-between the prediscussion attitudes in the unshared condition, whereas participants’ attitudes in the shared condition show the typical polarization effect (Vinokur & Burnstein, 1978). In light of this discussion, it seems that the distribution of stereotype-inconsistent information across the members of a target group may well have an impact on the degree to which group discussion affects subsequent stereotypes. But the direction of this moderation is unclear. If unshared information is not discussed, then greater polarization as a result of group discussion may happen when the inconsistent information is unshared rather than shared. On the other hand, that information is in some sense more novel in the unshared condition, and participants may actually be more likely to discuss it in that condition, inhibiting the polarizing effects of group discussion CURRENT RESEARCH ON THESE HYPOTHESES Study 1: Stereotype Acquisition Park and Hastie’s (1987) work provides convincing evidence for the idea that acquiring a stereotype from others leads to less variable perceptions of the target group than learning about the group through observing behavioral instances performed by members of the target group. Their research suffers, however, from a major drawback. The researchers constructed the abstract descriptions of the target group that were given to participants. Accordingly, we do not know whether those abstractions resemble the sort of abstractions about social groups that are communicated and that are actually the basis of abstraction-based stereotype formation outside of the experimental laboratory. In order to understand how abstractions of groups are actually formed and communicated, Thompson, Judd, and Park (2000) conducted a study in which they employed a rumor transmission paradigm. Participants were organized in groups of nine and linked in three triads of three persons within each nineperson group. Participants in triad 1 were asked to form an impression of a hypothetical fraternity by reading 45 behaviors performed by different individual members of the fraternity. These behaviors conveyed information along three different trait dimensions. In general they gave the impression that fraternity members were academically competent, espoused liberal political views, and assumed leadership roles on campus. After reading these behaviors, these triad 1 participants were asked to write a one-page description
182
THE PSYCHOLOGY OF GROUP PERCEPTION
of what the group was like. These participants then completed the dependent measures that assessed their impression of the group, including measures of perceived group variability and stereotype extremity. Participants in triads 2 and 3 formed an impression of the same group but they did so on the basis of the impressions that participants in earlier triads had written. Thus, triad 2 participants in each group of nine read the three-page-long impressions that triad 1 participants in their group had written and then wrote their own impressions. Triad 3 participants in each group read the three page-long impressions that triad 2 participants in their group had written. Additionally, groups of nine participants were randomly assigned to either an abstraction-only condition or an abstraction-plus-instance condition. Triad 2 and 3 participants in the former condition, after reading the written impressions from the three previous participants in their group, wrote their own page-long impressions and then completed the dependent measures. In the abstraction-plus-instance groups, triad 2 and 3 participants, after reading the group impressions written by prior participants in their group, read the 45 behavioral instances that had been seen by all triad 1 participants. They then wrote their own impressions and completed the dependent variables. Thus, the two conditions differed in whether triad 2 and 3 participants formed their impressions only on the basis of the previous triad’s impressions or on the basis of those impressions plus the behavioral instances. From the group ratings provided by participants, two different components of the group stereotype were assessed. The first consisted of what has elsewhere been labeled the perceived stereotypicality of the group (Park & Judd, 1990), i.e., the tendency to describe the group in stereotypically extreme ways or to ascribe many stereotypic traits and few counterstereotypic traits to the target group as a whole. The second component focused on perceived dis persion, i.e., the extent to which members of the target group are seen as being similar or dissimilar to each other. Multiple measures of each of these two components were assessed, all showing the same pattern of effects. For simplicity here we only report results on one measure of each. In the case of stereotypicality, we report estimates of the percentage of the group judged to have stereotypic versus counterstereotypic traits. In the case of dispersion we use an index computed as the range of trait judgments given to individual group members. Results on these two measures, broken down by triad and by condition, are shown in Figure 13.1. The measure of perceived stereotypicality (see top panel of Figure 13.1) showed a clear main effect of triad, F (2, 32)=6.93, p<.01. Later triads tended to see the target group more stereotypically than earlier triads. A closer look at the means makes clear that this main effect is mostly due to the difference between triad 1 on the one hand and triads 2 and 3 on the other hand. Thus, those who learned of the group from abstractions written by previous participants saw the group in a more extremely stereotypic manner than those who were only exposed to the behavioral instances from which those abstractions were composed. Differences attributable to further communication of an abstraction, between the second and third triad members, were not significant. Interestingly, the triad effect was not moderated by condition. Analysis of the perceived dispersion measure (see bottom panel of Figure 13.1) also revealed a significant main effect of triad, F (2, 32)=18.79, p<.001. As with stereotypicality, this effect is mostly due to the fact that participants in the first triad saw the target group as more dispersed than those in the second and third triad. The triad difference in dispersion was moderated by condition, however, F (1, 16)=8.99, p<. 01. The decrease in perceived dispersion as a function of the triad was less pronounced when triad 2 and 3 participants received both abstraction- and instance-based information than when they received only abstraction-based information. Thompson et al. (2000) also examined whether there was greater consensus in perceptions of the central tendency of the group in later triads than in triad 1 participants, where each person could form individual impressions from the behavioral data. They calculated intraclass correlations for each triad on measures of
THE ACQUISITION, TRANSMISSION, AND DISCUSSION OF SOCIAL STEREOTYPES
183
FIGURE 13.1. Perceived stereotypicality and dispersion of target groups by triad and by condition (Thompson, Judd, & Park, 2000, Study 1).
perceived stereotypicality. The results showed a significant increase in consensus from triad 1 to triads 2 and 3. Additional analyses revealed that this difference was not an artifact of the fact that all participants within a group read the exact same abstractions, since it was found even when participants across individual groups were mixed up. In other words, greater consensus in group stereotypes was found for participants who learned about the group through communications from others, even when the communications they received were not identical. Taken together, this first study provides convincing evidence for the first of the four hypotheses described above. According to this hypothesis, the way a stereotype is acquired influences the representation of the target group. Extending Park and Hastie (1987), individuals who learned about a group through secondhand communication, involving actual communications as opposed to experimenter-created ones, saw the group as significantly less variable than individuals who learned about the group through first-hand encounters. Unlike the Park and Hastie results, where the extremity of the group stereotypes was experimentally controlled, Thompson et al. (2000) found evidence that group impressions learned from abstracted communications are also more stereotypically extreme than instance-based stereotypes.
184
THE PSYCHOLOGY OF GROUP PERCEPTION
Study 2: Stereotype Transmission The purpose of Thompson et al.’s (2000) second study was twofold. On the one hand, they wanted to replicate the findings of the first study with a different experimental procedure. The abstractions in the first study were written one-page descriptions about the target group. Would stereotypes also be more extreme and homogeneous if they were acquired through direct verbal, rather than written, descriptions? Additionally, Thompson et al. wanted to examine the subsequent stereotypes of the communicator to see if the simple activity of telling other people about one’s impressions of a group causes the communicator to adopt a more polarized stereotype. Thompson and colleagues created four-person groups in the laboratory. Groups were randomly assigned to either a discussion or no-discussion condition. In the discussion condition, two of the four participants learned about one hypothetical target group by reading behavioral instances supposedly engaged in by group members, while the other two participants learned about a second target group in the same manner. They each then wrote summary impressions of the group they had learned about. Following this, all four participants came together with the goal of communicating their impression of the target group they encountered to the other two participants in the group. First one target group was presented, then the second, with order randomly determined. In each case the two participants who had earlier encountered the behavioral instances about that target group read their impressions of that group and then talked informally to the other two participants in order to convey their impressions more fully. Then the process was reversed, with the second two participants reading their written impressions of the target group they had encountered and presenting the group verbally to convey their impressions of this second group. All four participants then completed the dependent variables, providing stereotypicality and dispersion ratings for both target groups, with the order of group ratings randomized. Participants in the no-discussion condition first learned about one of the two target groups from the same behavioral instances as those used in the discussion condition. After writing an impression of the group, they then encountered the behavioral instances of the second target group and wrote an impression of this second group. Finally, these no-discussion participants completed the dependent measures for both groups. In sum, discussion versus no-discussion varied between participant groups. In the discussion condition, which target group was learned about from instances and which one was learned about from discussion varied within participant groups. Additionally, the order in which the two target groups were discussed varied between participant groups. In the no-discussion condition, all participants learned about both target groups from instances, but the order of exposure to this target group information varied between participants. As in Study 1, the dependent variables permitted the researchers to examine both the perceived stereotypicality and dispersion of the target groups. Again, a variety of measures of each of these constructs were used, all showing similar effects. As we did in Study 1, for reasons of efficiency we discuss results only for one measure of each, the percentage estimate measure of perceived stereotypicality and the range measure of perceived dispersion. The relevant means, broken down by condition and by order of learning, are shown in Table 13.1. The measure of perceived stereotypicality showed a main effect of condition, F (1, 35)=34.27, p<.001. On average, target groups were judged more stereotypically in the discussion condition than in the nodiscussion condition. However, there was no evidence that this condition main effect was further qualified by order of learning, F (1, 35)=0.36, p>.50. This null effect suggests that two very different activities have relatively similar consequences. Both the acquisition of a stereotype through a verbal description (rather than through contact with behavioral instances) and the activity of talking to other people about one’s impressions (rather than not talking) causes the stereotype to become more extreme.
THE ACQUISITION, TRANSMISSION, AND DISCUSSION OF SOCIAL STEREOTYPES
185
Analyses on the measure of perceived dispersion also revealed a significant main effect of condition, F (1, 35)=33.41, p<.001. This main effect was qualified by an interaction with order of learning, F (1, 35)=8.22, p<.01. The means in Table 13.1 suggest that those who saw the target group as most variable were the participants who read the behaviors, wrote their one-page description, and then filled out the dependent measures (i.e., no-discussion condition). Participants who learned about the target group only through the verbal report from others saw the target group as least variable. A final group of participants—those who read the behaviors, wrote a one-page description, and then communicated their impressions to other participants—fell in between. TABLE 13.1. Perceived Stereotypicality and Dispersion of Target Groups by Discussion Condition and Order of Learning (Thompson, Judd, &. Park, 2000, Study 2) Discussion
No Discussion
Measure
1st Target Group (Learned from instances, then transmitted to others)
2nd Target Group (Learned from verbal descriptions from others only)
1st Target Group (Learned from instances only)
2nd Target Group (Learned from instances only)
Perceived Stereotypicality Perceived Dispersion
82.68
82.26
70.36
64.60
10.61
9.22
13.76
14.05
In order to asses consensus, the authors calculated simple correlations between pairs of participants who encountered the particular target group that was being rated in the same order. These correlations were then analyzed as a function of condition and order of learning. A reliable main effect for condition revealed that there was greater consensus among groups in the discussion condition than among groups in the nodiscussion condition. This effect was not qualified by an interaction with order of learning, suggesting that the increase in consensus due to communication is equally strong for those who learn about the stereotype from others as for those whose task was to talk about their group impressions. Thompson et al.’s (2000) Study 2 nicely supports the first two hypotheses outlined in the introduction to this chapter. The basic results of Study 1 were replicated, this time with verbal instead of written descriptions: Learning about a target group through communication from others leads to more extreme and more homogenous perceptions of the target group than learning about the group through behavioral instances. Furthermore, Study 2 also showed that the activity of communicating impressions to others has the same effect. “Stereotype transmitters” saw the target group more stereotypically than participants in the no-discussion condition who had received the same information but who did not have the chance to talk about their impressions. Study 3: Discussion and Information Pooling Brauer, Judd, and Jacquelin (2001) conducted two studies that were designed to test the third and fourth hypotheses described above, extending the Thompson et al. (2000) results. First, the communication in the Thompson et al. studies already reviewed was formally constrained by the experimenters so that it was more of a one-sided communication from one person or persons to others than a real group discussion. Brauer et al. sought to demonstrate similar effects on stereotypicality and perceived group dispersion using real group discussions. Secondly, the studies reported by Brauer et al. examined whether these group discussion effects would be moderated by the distribution of counterstereotypic information in the discussion groups.
186
THE PSYCHOLOGY OF GROUP PERCEPTION
Brauer et al. (2001, Study 1) gave groups of three participants information about a hypothetical target group. The information took the form of 114 different behaviors that were written on index cards and that had presumably been performed by members of the target group. Of these, 18 behaviors were neutral (both evaluatively and in terms of the stereotypic content of the group). The remaining 96 behaviors referred either to the dimension of selfishness or cowardliness. The majority of these (75%) described selfish and cowardly behaviors. The remainder (25%) described behaviors that were inconsistent with this general stereotype, i.e., they described altruistic and courageous behaviors. Thus, the group was stereotypically described as selfish and cowardly, but there were numerous counterstereotypic behaviors. The distribution of the different behaviors across group members was varied, so that the counterstereotypic information was shared in half of the groups and unshared in the other half of the groups. In the shared condition, each group member read 12 stereotypic and 4 counterstereotypic behaviors on each of the two trait dimensions used to describe the target group. In the unshared condition, one group member saw the majority of the counterstereotypic behaviors (8 on each trait dimension) while the other two group members saw relatively few of them (2 on each trait dimension). The total number of behaviors read by each group member was held constant, so that in the unshared condition, the participant who saw more counterstereotypic behaviors saw fewer stereotypic ones. The other between-group manipulation varied whether or not there was group discussion. Half of the groups were asked to discuss their impressions about the target group and then filled out dependent measures assessing perceived dispersion, perceived stereotypicality, and liking for the target group. The other half of the groups did not engage in a group discussion and filled out the dependent measures immediately after having read the behaviors. As before, multiple measures of perceived stereotypicality and dispersion were gathered, and within each set they all showed consistent results. Therefore, we present results only for the percentage estimate measure of stereotypicality and the range measure of dispersion. Additionally, liking for the target group constituted a third dependent variable. These variables were analyzed as a function of condition (discussion versus no discussion) and distribution of the counterstereotypic information (shared versus unshared). The relevant means for the perceived stereotypicality ratings are reported in Figure 13.2. The analyses confirmed the authors’ hypotheses, albeit only for one of two trait dimensions (selfishness) and not for the other (cowardliness). For the measure of perceived stereotypicality on selfishness, there was a main effect for discussion, F (1, 41)=4.62, p<.05 (see top panel of Figure 13.2). This main effect was qualified by an interaction with information distribution, F (1, 41)=5.09, p<.05. The informal group discussion led to more stereotypical perceptions of the target group, but only when the counterstereotypic information was shared rather than unshared. No reliable effects emerged on the cowardliness trait dimension (see bottom panel of Figure 13.2). Additional analyses indicated that this may have been due to a failure to induce a target group stereotype along this dimension. Unlike the selfishness dimension, where the group was clearly seen as being selfish rather than altruistic, the target group was not judged to be either particularly cowardly or particularly courageous. Thus, the behaviors we presented did not convey the stereotypic impression of the target group that we meant to convey.1 It hardly seems surprising that group discussion would fail to induce greater perceived stereotypicality of the target group when the group was not seen in a stereotypic manner on the cowardly trait dimension in the first place. The perceived dispersion measure was unaffected by any of the independent variables, all Fs<1. Analyses on the liking measure yielded a significant discussion by information distribution interaction, F (1, 41)=4. 08, p<.05. Groups in the discussion/shared condition liked the target group significantly less than groups in the other three conditions.
THE ACQUISITION, TRANSMISSION, AND DISCUSSION OF SOCIAL STEREOTYPES
187
FIGURE 13.2. Perceived stereotypicality on the egoism dimension (top panel) and the cowardly dimension (bottom panel) as a function of discussion condition and information distribution (Brauer, Judd, & Jacquelin, 2001, Study 1).
In sum, Brauer et al. (2001) found evidence for the polarizing effect of communication that had also been demonstrated by the Thompson et al. (2000) studies, but only on one of the two trait dimensions. This time, however, the communication consisted of an actual group discussion of the target group rather than a structured one-way communication from one set of participants to others. Additionally, and importantly, this polarization occurred only in discussion groups where the counterstereotypic information was shared by all members of the discussion group. In the condition where the majority of the counter-stereotypic information was possessed by only one of the discussion group members, no polarization was found. Finally (unlike the Thompson et al. results), no effects on the perceived dispersion of the target group were found. Because of the failure to induce the desired group stereotype on the second trait dimension, and the resulting absence of discussion effects on this dimension, Brauer et al. (2001) conducted an additional study to replicate these results with different experimental stimuli. This replication additionally seems warranted in light of two considerations. First, unlike the Thompson et al. (2000) results, no discussion effects on perceived target group dispersion were found. Second, although the distribution of the counterstereotypic information across discussion group members was found to moderate the impact of group discussion, initial predictions about the direction of this moderation were not strongly made. Accordingly, more processoriented measures were assessed to examine the mechanisms underlying the obtained moderation.
188
THE PSYCHOLOGY OF GROUP PERCEPTION
Study 4: Discussion as a Mediator As just discussed, Brauer et al. (2001) conducted a second study to replicate the effects they found in their first study and in order to better understand the mechanisms underlying these communication effects. The procedure of their second study resembles that of the first one but with a number of important modifications. First, the cowardly and courageous behaviors were replaced by violent and nonviolent behaviors. Given the distribution of behaviors—there were three times as many selfish and violent behaviors as there were altruistic and nonviolent behaviors—the target groups stereotype was that it was relatively selfish and violent. Second, all the group discussions were audio-taped, transcribed, and coded in order to assess what group members talked about during the discussion. Third, participants were given a surprise recall test at the end of the experiment that allowed the authors to examine the relationship between perceived variability and memory for stereotypic and counterstereotypic behaviors. TABLE 13.2. Perceived Stereotypicality and Dispersion of Target Groups as Well as Liking for Target Groups by Discussion Condition and Information Distribution (Brauer, Judd, & Jacquelin, 2001, Study 2) Shared Condition Measure and Dimension Perceived Stereotypicality Selfishness Cowardliness Perceived Dispersion Selfishness Cowardliness Liking Temperature Rating
Unshared Condition
No Discussion
Discussion
No Discussion
Discussion
31.94 26.36
54.26 48.86
35.01 29.94
34.91 31.89
19.50 18.07
17.61 18.00
18.45 18.89
18.34 17.47
42.75
32.50
42.50
45.59
Mean stereotypicality scores, dispersion scores, and liking scores are shown in Table 13.2. This time, the main effect of discussion on perceived stereotypicality of the target group was significant for both trait dimensions, selfishness: F (1, 60)=6.04, p<.05, violence, F (1, 60)=6.95, p<.05. Discussion, however, polarized stereotypes more in the shared condition than in the unshared condition, as evidenced by a significant condition by information distribution interaction, F (1, 60)=6.15, p<.05 (selfishness) and F (1, 60)=4.90, p<.05 (violence). As in the previous study, perceived dispersion was not affected by any of the independent variables. Analyses of the liking scores revealed a significant interaction between discussion and information distribution, F (1, 60)=4.52, p<.05. Again, participants in the discussion/shared condition felt less warmly toward the target groups than participants in the other three conditions. The mean values of the second stereotypicality measure for each of the three group members separately are graphed in Figure 13.3. The responses of participants in the shared condition are shown in the left half of the figure. Remember that in this condition, members A, B, and C received the same proportion of stereotypical and counterstereotypical behaviors. Variation among the three group members can thus be treated as random error, and this random error has been removed in the figure. The responses of participants in the unshared condition are shown in the right half of Figure 13.3. In this condition, it was always member C who received a large proportion of counterstereotypic behaviors whereas members A and B mostly read stereotypic behaviors. For the reasons stated above, differences between members A and B (which can be considered random variation) have been removed in the figure.
THE ACQUISITION, TRANSMISSION, AND DISCUSSION OF SOCIAL STEREOTYPES
189
FIGURE 13.3. Perceived stereotypicality for each of the three group members, A, B, and C, as a function of discussion condition and information distribution (Brauer, Judd, & Jacquelin, 2001, Study 2).
Figure 13.3 gives some insight into how group members’ perceptions shift as a result of group discussion. In the shared condition, group members see the target group as being moderately stereotypic if they give their ratings right after having read the behavioral information. However, those who had the chance to discuss their impressions with others tend to adopt relatively polarized attitudes toward the target group, i.e., they see the target group as being very stereotypical. It seems, then, that discussion causes group members to shift toward a position that is more extreme than their average position prior to the discussion. The situation is quite different in the unshared condition. Here, group members A and B see the target group as very stereotypical and member C does not see the target group as stereotypical at all if their ratings are based on the behavioral information only. In comparison, members A and B’s ratings are less polarized and member C’s ratings are more polarized if the three group members have a chance to exchange their viewpoints with each other. It seems as if discussion causes group members to converge on a position that is a compromise of their positions before the group discussion. In other words, we find group polarization in the shared condition but convergence on the average position in the unshared condition. Independent coders who were blind to condition listened to the tape-recorded discussions of the groups in the discussion condition and measured the amount of time that each group talked about stereotypic and counterstereotypic behaviors. Not surprisingly, all groups talked more about stereotypic than about counterstereotypic behaviors, F (1, 28)=7.35, p<.01. But this difference depended on the distribution of information in the group. In the unshared condition, the difference in time spent talking about stereotypic versus counter-stereotypic behaviors was significantly less than the difference in the shared condition, F (1, 28)=4.80, p<.05. Mediational analyses revealed that the effect of information distribution on perceived stereotypicality was mediated by the discussion content: Groups in the unshared condition talked more about counterstereotypic and less about stereotypic behaviors than did groups in the shared condition, and this difference in turn affected the perceived stereotypicality ratings of the group members after the discussion. The surprise recall test administered at the end of the study yielded a number of interesting results. Participants who discussed their impressions of the target group had a recall advantage over participants in the no-discussion condition, F (1, 60)=34.21, p<.001. Participants in the discussion condition recalled counterstereotypic behaviors better than stereotypic behaviors, but this difference was due entirely to one cell in the experimental design, namely, participants the discussion/unshared condition. All other participants recalled roughly an equal proportion of stereotypic and counterstereotypic behaviors. The triple interaction
190
THE PSYCHOLOGY OF GROUP PERCEPTION
between condition, information distribution, and behavior type (stereotypic versus counterstereotypic) was statistically significant, F (1, 60)=11.16, p<.001. It should be mentioned that the recall measures did not correlate with the perceived stereotypicality measure. Memory for behavioral instances was unrelated to the overall judgment that participants made about the target group as a whole. Such a null result is not rare when partici pants are given instructions to form an impression and make judgments about a target on-line while they process the relevant information (Hastie & Park, 1986). THEORETICAL INTEGRATION AND CONCLUSIONS These four studies tell a remarkably clear and consistent story about the impact of communication on social stereotypes. And that story helps us refine our conception of stereotypes as social constructions, i.e., as shared stories about social categories that we tell each other and that come to define our social world. We quickly review the conclusions from these studies, attempting to integrate them and provide a framework for understanding their diverse results. The most consistent conclusion across all four studies is that the communication of stereotypes affects the extremity of stereotypic beliefs about social categories. In Study 1, individuals who heard about target groups from others held more extremely stereotypic views of those groups than did the individuals who formed their views based on behavioral observations. This difference in stereotype extremity persisted even when those who learned about the group via second-hand impressions also were exposed to the behavioral observations that formed the basis for the group stereotype in the first place. Thus, even given equivalent behavioral observations of the target group in question, stereotyped communications about that group from others increases the extremity of the subsequent group perceptions. The second study showed that stereotype extremity is affected not only by being the recipient of a communication, but also by formulating and transmitting that communication. Compared to those who simply studied the behavioral material, those who told others what their group impressions were like came away with more stereotypic views of the target groups. And the third and fourth studies showed that both of these one-way communication effects combine in some sense in real group discussions, where everyone is both a transmitter of impressions and a recipient of others’ target group impressions. A major qualification to these conclusions concerning the impact of communication on stereotype extremity derives from the manipulation of information sharing in the third and fourth studies. Consistent with what one might have expected from the group polarization literature, the group discussion effects on stereotype extremity were consistently found only when all participants in the group shared more or less the same initial target group stereotypes. In this case, the group discussion focused largely on stereotypic information and subsequent stereotypes were more extreme. On the other hand, when one of the discussion group members initial view of the target group was rather different, due to the fact that this member had learned about considerably more counterstereotypic information, then the largely unshared counterstereotypic information was the subject of greater group discussion and polarization of target group stereotypes did not ensue. Although the communication effects that these studies document on stereotype extremity seem quite consistent, those pertaining to the impact of communication on perceived group dispersion, or within-group heterogeneity, might initially seem to show less consistency. But in fact we believe that these results are quite consistent although the story here is a bit different than the story for stereotype extremity. In the first two studies, individuals who were exposed to communication from others about a target group subsequently saw the target group as less dispersed or variable, but only if they had not previously been exposed to the individuating behavioral instances. In the third and fourth studies, no effects on dispersion were found, but here, too, all participants had previously seen a range of behavioral instances from the target group. Thus it seems that the most parsimonious conclusion that can be taken away from these results is that stereotypes
THE ACQUISITION, TRANSMISSION, AND DISCUSSION OF SOCIAL STEREOTYPES
191
that result from communicated abstractions about the group in fact do contain less within-group variability information (consistent with Park and Hastie, 1987), but that prior acquaintance with individuating group instances can mitigate this effect. In this regard, it is perhaps useful to speculate a bit about the various measures used to assess withingroup variability or dispersion. The two measures routinely used include a range measure (the task from which results reported in this chapter were computed), in which participants rate the highest and lowest group members on trait attributes, and a histogram task, in which participants divide the group up into a subjective frequency distribution. Both of these tasks require that the participant, to some extent, think about the group at the instance or exemplar level, rather than at the level of the group as a whole. Therefore, it seems to us not surprising that individuals who receive communications from others after learning about the individual group instances can still bring to mind those diverse individual instances, even though the group stereotype has been altered through the communication process. Over time, of course, in the absence of further direct contact with individual group members, the superordinate group stereotype, derived from and reinforced by communication with others, should come to dominate. As already mentioned, the results of the information distribution manipulation in the third and fourth studies discussed in this chapter suggest that the polarizing effects of communication and group discussion on stereotypes can be inhibited if there exist differences in the initial beliefs about the target group. In the case we explored, one discussion group member had been exposed to much more counterstereotypic information about the target group than the other discussion group members. Interestingly, and contrary to work that has tended to argue that information that is unshared among group members tends not to be discussed (Gigone & Hastie, 1993, 1997; Stasser, Taylor, & Hanna, 1989; Winquist & Larson, 1998; Wittenbaum & Stasser, 1998), counterstereotypic information was more likely to be discussed when it was largely unshared than when it was shared. It is important to note here that, strictly speaking, all individual pieces of information were unshared, since only one participant saw each individual behavioral instance. However, in some ways the stereotypic instances were all highly redundant with each other in the information they conveyed about the group, as were the counterstereotypic behavioral instances. Thus, it was the type of information, rather than the specific instances themselves, that was shared versus unshared according to the information distribution manipulation. It seems that the counterstereotypic information was regarded as particularly informative or diagnostic when it was unshared. It was discussed at greater length, and this discussion seemed to inhibit the polarizing effects of discussion on subsequent stereotypes. Stereotypes are often referred to as “social” stereotypes. The usual sense of this adjective is that they refer to sets of beliefs about “social” categories, groups defined by ethnicity, gender, religious beliefs, and a host of other “social” markers that we use to categorize our world. We hope that the research we have discussed in this chapter illustrates that there is another meaning to the “social” modifier that has not to date been emphasized in the literature but that has important theoretical and practical consequences. Stereotypes are “social” in the sense that they are social products or constructions that reside not solely in the minds of individual perceivers but that develop and operate through interpersonal interaction and communication (S.A.Haslam, Turner, Oakes, McGarty, & Reynolds, 1998;S.A.Haslam, Turner, Oakes, Reynolds, Eggins, Nolan, & Tweedie, 1998). And we believe that such communication has the potential to profoundly influence the social stereotypes that are prevalent in our society. Through communications, stereotypes in some sense acquire a life of their own, perhaps divorced from any empirical reality about actual group differences. Communication, we believe, is largely responsible for the fact that stereotypes do not accurately reflect the diversity of the social world. Rather they tend to be shared exaggerations or caricatures of that social world.
192
THE PSYCHOLOGY OF GROUP PERCEPTION
NOTE 1. The target group was described to participants as a group of campers. We presumed that they would be seen to be about 15 years old, and the cowardly behaviors were written with such an age group in mind. In fact, participants reported that they thought the typical camper was 10 years old. Our manipulation failure was due to this difference, we believe, because behaviors that are perhaps cowardly for a 15-year-old probably are not cowardly for a 10-year-old target
14 Culture, Communication, and Entitativity A Social Psychological Investigation of Social Reality YOSHIHISA KASHIMA Department of Psychology, The University of Melbourne
We humans live in social reality that we ourselves have constructed. From marriage and money to nationstates, the European Union, and the United Nations, we live in the reality that has been constructed by humans in the past and is being maintained by us in the ongoing activities of our everyday lives. In fact, social scientists of the past and present took it for granted that humans construct social reality. For example, the classical sociology of Durkheim (1901/1982) and a more recent rendition of it by Giddens (1976/1993) both took the analysis of social institutions as a central question of sociology (and social sciences). The difference between them mainly lies in their assumption about the extent to which social reality is given (Durkheim) or reproduced in social activities (Giddens). Regardless of the way in which a group of individuals constructs social reality, once constructed, it becomes part of what they take for granted, and thus constrains and enables what they can and may do in the future. Arguably, the main point of the social sciences is to examine the working of social reality from a variety of perspectives. In experimental social psychology, too, social interactions in small groups were regarded as constructing a social reality that is largely emergent in the particular situation (e.g., Asch, 1952; Festinger, 1950; Sherif, 1936). Yet the concept of social reality went out of favor as the discipline turned cognitive when the cognitive revolution of the 1960s put cognitive consistency and attribution research, and more generally social cognition, at the center of the disciplinary research agenda. Despite notable exceptions (e.g., Moscovici, 1976; Ostrom, 1984; Rommetveit, 1974; Tajfel, 1981b; Turner, 1991), the critical observation that social interaction constructs social reality and social psychological processes operate within it had lain dormant in the background of social psychological theorizing. Nevertheless, a recent return of interest in the social basis of cognition (e.g., Hardin &: Higgins, 1996; Levine, Resnick, & Higgins, 1993) brought forward the question of social reality, that is, how social reality is constructed and maintained, and what psychological processes are involved in its production and reproduction.
194
THE PSYCHOLOGY OF GROUP PERCEPTION
The main purpose of this chapter is to place the concept of social reality within a social cognitive framework that takes culture and communication seriously. In particular, my aims are twofold. First, I wish to sketch out a theoretical framework in which to consider relationships among social reality, culture, and communication. In particular, I argue that communication plays a significant role in the production and reproduction of culture and social reality by generating and maintaining the psychological sense of its realness. Second, I wish to make a case that social groups, not only small interacting groups but also social categories, constitute significant part of social reality for humans, and the concept of entitativity is centrally involved in this social ontology. In particular, I argue that (1) Campbell’s concept of entitativity is meant to capture how real a human aggregate is, and that (2) when people have a psychological sense of the realness of social categories (i.e., perceived to be entitative), they tend to apply psychological essentialism to these collectives (i.e., members of the social categories are seen to possess essences). Note that the foregoing arguments imply close conceptual relationships among culture, communication, and the psychological sense of the realness of social groups. I briefly review the literature to support these claims. CULTURE, COMMUNICATION, AND SOCIAL REALITY How is social reality constructed, and how are culture and communication involved in its construction? To answer the “how” question, it is necessary to examine preliminarily the “what” question of social reality and consider the nature of social reality that we regard as peculiarly human (e.g., money, nation-states). Just such a treatment was recently given by a well-known philosopher of language and mind, John Searle (1995, 1998). Although a full treatment of his theory goes far beyond the scope of this chapter, I believe it is correct in its general outline. At the very least, his theory can provide a general framework to sketch out the relationships among culture, communication, and social reality. Searle’s Ontology of Social Reality Searle argued that the intentional human mind and its symbolic capacity provide the basis for the construction of social reality. Humans can have collective intentionality, in which a group of people shares a belief, desire, or intention that “we,” that is, the group, are engaged in certain activities together. Searle uses here the term intentionality in a technical philosophical sense, where it includes not only intention, but also belief, fear, and want, any mental state that is about some object or proposition. Therefore, one may hold that “I intend to play soccer,” or “We intend to play soccer.” The former indicates singular intentionality and the latter collective intentionality. According to Searle, when a collection of individuals shares collective intentionality, this constitutes a social fact, that is, “playing soccer.” He contends that collective intentionality may occur, and therefore a social fact may be constructed, by any social animals when they are engaged in coordinated activities (e.g., a pack of hyenas hunting for a lion). What makes some of the human social facts peculiarly human is that we construct institutional facts. According to Searle, institutional facts are constructed by collectively imposing (that is, by collective intentionality) what he calls a status function on brute facts. To see how this works, consider the example of Australian bank notes. They are printed on a polymer (plastic) substrate rather than on paper as in many other countries in the world. A $10 note in Australia is a piece of plastic (a brute fact), but it is assigned a function of “money” by a rule that says, roughly, “This type of piece of plastic counts as money in the country where it was made.” In other words, this rule imposes a certain status (“$10 note”) on a brute fact (piece of plastic). Such an assignment of status to a brute fact constructs an institutional fact, which has a function that cannot be achieved by the brute fact’s physical and chemical makeup.
CULTURE, COMMUNICATION, AND ENTITATIVITY
195
Searle notes that underlying the status function is humans’ symbolic capacity, in which an object is used to symbolize or represent something else. That is, it is possible for humans to think and act so as to imply that an object X counts as Y in context C. Searle calls the rule of the form “X counts as Y in context C” a constitutive rule. Note that “This piece of plastic counts as money in Australia” is a special case of this general form “X counts as Y in context C.” Note that this rule specifies a symbolic relationship between the piece of plastic and the social function of “money,” in that the plastic symbolizes money. Without the human capacity to symbolize (i.e., to represent something by something else), the brute fact of the plastic cannot be transformed into the institutional fact of money. To Searle, what is fundamental to the construction of an institutional fact is the human capacity for symbol use. Although the ontology of social reality may be what Searle says it is (and there may very well be much philosophical debate here), his theory leaves much to be desired from social psychological viewpoints. Overall, Searle’s theory is concerned with the conceptual structure of social reality; i.e., what ingredients are needed and how they are conceptually related to each other. However, this lacks social psychological dynamics in time and space. Questions pertinent to social psychologists have more to do with the process of constructing social reality, that is, how we construct social reality in our ongoing social activities. What follows is an outline of a social psychological theory of the construction of social reality. Culture, Communication, and the Psychological Sense of Social Reality Central to Searle’s ontology of social reality is the human capacity for symbol use and collective intentionality. However, in order to theorize how people construct social reality, a series of fundamental questions needs to be addressed. First, although humans may have a capacity for symbol use, we must have symbols to begin with. Where, then, do symbols come from? Second, Searle takes for granted that humans are capable of holding shared collective intentionality. However, for social psychologists, a fundamental question is how we come to hold a shared collective intentionality. I contend that answers lie in culture and communication. First of all, symbols come from culture. When symbols are broadly conceptualized as including language and other types of things that refer to something else, a collection of symbols is a large portion of what we often call culture. In this sense, culture is a collection of symbols that have been cumulated over time within a group of people (see Kashima, 2000a, in press, for brief discussions on this). For example, the cultural symbol “money” and a network of other cultural symbols and practices associated with “money” in Australia (e.g., legislation, rules, and regulations of the Reserve Bank of Australia, production procedures at the Royal Australian Mint) provide symbolic resources that embody the constitutive rule “This type of plastic counts as a $10 note in Australia.” This constitutive rule constructs social reality, or an institutional fact, of the $10 note. In this sense, the $10 note is “an entity created by the social agreement that something counts as that entity (D’Andrade, 1984, p. 91; emphasis in the original).” In other words, cultural symbols and practices help people construct social reality by embodying constitutive rules. In turn, the social reality thus created provides referents that those symbols refer to, that is, a real money. Culture and social reality therefore are conceptually bound together; that is, the culturally constituted symbols and their referring social reality constitute an interlocking system. Let us turn to the second question. How do we come to hold shared collective intentionality? To give a more concrete example, how does each member of a group come to hold that “We believe this plastic counts as money in Australia?” We achieve this through communication, that is, communication under stood as the socially coordinated process of establishing a mutual understanding among communicators. Clark and his associates (1996; Clark & Wilks-Gibbs, 1986; Clark & Brennan, 1991; for a review, see
196
THE PSYCHOLOGY OF GROUP PERCEPTION
McIntyre, Lyons, Clark, & Kashima, in press) called this process grounding. According to Clark, language use and communication in general involve socially coordinated activities in which communicators cooperate to assign symbols (e.g., particular phrases, expressions, and sentences in a natural language, or other forms) to their referents. So, for instance, initially, in an attempt to refer to a complex figure that cannot be described by one word, a communicator may describe it as “the figure that looks like a skater sticking his leg to the right.” Clark calls this type of verbal communication a presentation. Before the communicator proceeds further in this communication, the audience would need to make an acceptance, that is, to show a sign that it has understood what the communicator means. If the audience fails to show some sign that he or she understands what is meant by a presentation, the communicator will make another attempt at presenting the same idea in a different manner. Nevertheless, as the communicator and audience become used to this type of communication, they will often establish a local convention (usually implicitly, but sometimes explicitly) to refer to figures by short phrases such as “the skater.” Clark’s research program illustrates several significant points. First, a communicator makes use of a natural language, English, as a common symbolic resource (or culture) shared with the audience to describe the figure verbally. To this extent, a large-scale culture participates in this communicative activity as a significant resource. Second, communication is a socially coordinated activity in that both the communicator and audience have to cooperate at one level to establish a mutual understanding. Third, when a local convention is established to refer to the figure as “the skater,” it is possible to regard this as a kind of constitutive rule, X counts as Y in context C, where X is the symbol token, “the skater”, Y is the figure, and C is the context of conversation between this communicator and this audience. To the extent that the same communicator keeps using the same phrase to refer to the same figure, it reproduces the convention in this context, thus reproducing a miniculture. Finally, there are at least two elements that are grounded in this communicative activity. One is the referential relationship between the symbol, “the skater” and the figure. The other is the collective intentionality that it is “we” (i.e., this communicator and this audience) who recognize and use this referential phrase. Thus, communicative activities construct and maintain the triadic relationship among publicly shared symbols (e.g., “the skater”), social reality as their referent (e.g., the figure), and collective intentionality of the referential relationship in the minds of the participants of the communication (e.g., “We know that ‘the skater’ refers to the figure”). In addition to the above issues of culture and communication, I believe that Searle’s investigation leaves out a question that is fundamental to social psychology. That is, where does the psychological counterpart to social reality come from? This is the question about the psychological sense of social reality that Festinger (1950) called subjective validity. What I mean by the psychological sense of social reality is best illustrated by considering its absence—the sense of social unreality. International travelers often experience this when they are faced with institutional facts such as foreign currencies that they have never seen before. As noted before, Australian banknotes are currently made of a flexible plastic substance. Visitors from overseas, where notes are typically made of paper, initially feel the psychological sense of unreality and incredulity, and feel as if it is a toy money though they are certain at one level that those notes are real money in Australia. However, as the foreign visitors use them in everyday transaction, they eventually gain the psychological sense that they are in fact real money. Where does this psychological feeling of reality come from? Again, an answer seems to lie in communication. When communicative activities construct and maintain collective representations that bind symbols with social reality, echoing Festinger (1950) and more recently Hardin and Higgins (1996), I believe that the psychological sense of the social reality is also strengthened by ongoing mutual verification of the collective intentionality. The psychological process behind this may have been captured by Higgins and his associates’ research (e.g., E.T.Higgins & Rholes, 1978; E.T.Higgins,
CULTURE, COMMUNICATION, AND ENTITATIVITY
197
FIGURE 14.1. The triadic relationship among culture, communication, and social reality and the placement of the concepts of entitativity and perceived entitativity of groups within the framework
1992) on “saying is believing.” This research found that when people described a target person to different audiences, their person descriptions changed as a function of the relationship between the audience and the target person. Later on, however, people ended up believing their own person descriptions despite the original information. It is important to note, however, when people communicate, what is conveyed is more than what is said explicitly. For example, saying “The Australian Labor Party pays lip service to environmental protection” presupposes that there exists an Australian Labor Party. This is an example of an existential presupposition (e.g., Green, 1996). More generally, in saying that “X is like Y”, the communicator implies the existence of X. If the “saying is believing” phenomenon captures the creation of meaning as Higgins (1992) claims it does, people may end up believing not only what they said explicitly, but also what they implied by saying it, namely, the existence of X. Indeed, in their discussion of the construction and maintenance of social reality, Berger and Luckmann (1966/1971) pointed out the importance of implicit presuppositions in utterances that are exchanged in routine conversations: Most conversation does not in so many words define the nature of the world. Rather, it takes place against the background of a world that is silently taken for granted. Thus an exchange such as, “Well, it’s time for me to get to the station,” and “Fine, darling, have a good day at the office,” implies an entire world within which these apparently simple propositions make sense. By virtue of this implication the exchange confirms the subjective reality of this world, (p. 172; emphasis in the original) To sum up, culture, communication, and social reality constitute a triadic engine of meaning making (see Figure 14.1). Culture supplies shared symbols that enable people to construct their social reality. Social reality then underpins the referent of the cultural symbols, which helped to construct the social reality to begin with. Nevertheless, it is the communicative acts of concrete individuals that get this process going.
198
THE PSYCHOLOGY OF GROUP PERCEPTION
People construct what Searle called a social fact by communication, by establishing collective intentionality among those who are engaged in communication. Once created, in communicating about social reality, people not only renew the social reality itself, but also enhance the psychological sense of the social reality by mutually verifying its very existence. The process of mutual verification of the collective intentionality about the existence of the social reality (Hardin & Higgins, 1996, called this shared reality) generates and maintains the psychological sense of the social reality, which in turn may further motivate people to reproduce the social reality. SOCIAL REALITY, GROUPS, AND ENTITATIVITY Apart from the individuals and institutional facts of society, what else constitutes our social reality? The strong methodological individualism that permeated social psychology often held that what constitutes social reality is individuals, not groups. Groups were regarded as “mere” aggregates of individuals (e.g., F.H.Allport, 1924). Whereas there were some voices of dissent (e.g., Asch, 1952; Sherif, 1933), together with the strong culturally shared individualist conception of the person (see Kashima, in press, for a review), the view that groups are real existences in the world has been regarded with suspicion. However, there is good evolutionary reason for suspecting that some groups may constitute a significant part of human social reality. It is well known that early hominids moved in bands. For evolving Homo sapiens, such bands were a constant feature of the species-specific environment to which it needed to adapt (Caporael, 1997). Dunbar’s (1996) hypothesis is that the human brain structure (especially the neocortex) evolved so that it enabled humans to have the kind of social coordination that we are so clearly capable of engaging in. As Hirschfeld (1996) argued, there maybe some genetically inheritable mechanisms that may help humans to regard groups as part of our social reality. In fact, there is a tradition of small group research (e.g., McGrath, 1984) and a tradition of social identity research (e.g., Oakes, Turner, & Haslam, 1994; Turner, 1987a) that regards a group as a real existence, that is, as part of our social ontology. In particular, the central theoretical question here is how some social aggregates are seen to be more real than others. An example may help clarify this. Consider an aggregate of people who live between 140 and 150 degrees east longitude. It is a social aggregate or what may be called a social category that is unlikely to be seen as real. This is presumably because the category boundary defined by longitude seems arbitrary. Nonetheless, under some circumstances such a social aggregate may come to be seen as a real entity. Imagine that a meteor is predicted to hit earth between the 140 and 150 degrees east longitude. Might not this knowledge of common fate confer on this seemingly arbitrary category of people a real ontological existence? If in fact some social aggregates seem arbitrarily defined, and yet are perceived to be real under some circumstances (Haslam, Rothschild, & Ernst, 2000), then what effects this psychological transformation? Entitativity, I believe, is a critical concept in this theoretical development. Entitativity: Social Ontology of Human Groups Campbell’s (1958) concept of entitativity is critically concerned with the ontology of social groups. Because there is a great deal of confusion about the term entitativity, some exegesis may help. In his enigmatic paper, Campbell’s objective was to give a scientifically satisfactory approach to determining whether an aggregate of persons is an entity or not. He specified the question with which he was to be concerned in these words: “Among actual or potential aggregates of persons, there are certain aggregates which meet criteria of being ‘entities,’ and other aggregates which do not. The distinction is capable of empirical representation” (p. 15). Campbell cited notable social scientists before him and called attention to their “common preoccupation
CULTURE, COMMUNICATION, AND ENTITATIVITY
199
with the ‘reality’ of social units” (p. 17). Without affirming or denying the reality of social aggregates as entities, he pointed out that these writers “propose [d] empirical tests of the ‘reality’ of social units” (p. 17), and went on to propose that “we look to the empirical clues of entity used in the visual perception of middlesized physical entities and then employ these clues in the analysis of social aggregates as entities” (p. 17). Foreshadowing the stance of evolutionary epistemology, which he was to develop later (e.g., Campbell, 1987), Campbell justified his move by arguing that the evolutionary process endowed us with a visual capacity that enables us to “analyze…entitativity so rapidly and vividly” (p. 17). Presumably, Campbell surmized that epistemic activities closely tied to evolution (i.e., organs of perception) would provide more valid or reliable clues to an empirical determination of reality. Campbell’s definition of entitativity appears in a footnote, as if it were a dictionary entry. It reads: “Entitativity: the degree of being entitative. The degree of having the nature of an entity or having real existence” (p. 17). Clearly, by this term he meant to refer to the ontology of a social group, that is, whether a human aggregate is a real entity or not. His discussions about perceptions of entitativity signify his attempt at empirically measuring it. His celebrated enumeration of Gestalt principles, such as proximity, similarity, and common fate, as determinants of perceived entitativity comprised the then available set of perceptual clues that he thought would afford our “vivid and rapid” and “seemingly direct” experience of the realness of entities. Campbell’s point was that, given the size and evolutionary makeup of humans, we are more attuned to perceiving entities of the size of human bodies and thereabouts (e.g., rocks and trees) but not of the size of human aggregates of thousands or millions. Nonetheless, his best bet was that, had we had an epistemic process that enabled us to perceive the entitativity of things of larger size, we might have used clues similar to those used by the visual mechanism. So the term entitativity was meant to capture the extent to which an aggregate of humans was a real entity; Campbell’s empiricism required that it be measured in some way. Psychological Essentialism as Perceived Entitativity of Social Categories If Campbell meant by group entitativity the ontology of human aggregates, what might be the psychological counterpart to the realness of a human aggregate? That is, what are the indices of our psychological sense of whether a human aggregate is a real entity or not? With Rothbart and Taylor (1992) and Yzerbyt and his colleagues (Yzerbyt, Rocher, & Schadron, 1997; Yzerbyt, Rogier, & Fiske, 1998), I believe that psychological essentialism (Medin & Ortony, 1989) may reflect people’s psychological sense of realness. Simply put, my postulate is this: When a human aggregate such as a nation, women, Jews, or blacks is regarded as having a real existence or having high entitativity, people may believe that the category may possess an essence, a set of properties that are unalterable and defining, that makes the social category what it is. To this extent, when a social category is believed to possess an immutable essence, it may be taken as evidence that people believe in its real existence. This line of reasoning derives from the research on children’s acquisition of natural-kind and artifact categories. It is well known that even young children can distinguish natural-kind categories (e.g., skunk, raccoon) and artifacts (e.g., chair, table; for reviews, see Gelman & Coley, 1991; Keil, 1989). For instance, children can tell that an artifact such as a coffeepot can be altered to a bird feeder, but a natural kind like a raccoon cannot be surgically made into a skunk, even if it has stripes and smells like the latter (Keil, 1989). Put differently, people appear to believe that a natural kind cannot be made into another natural kind by human manipulation but that an artifact can be. The immutability of a natural kind may stem from psychological essentialism, that is, the folk belief that there is an underlying essence to a natural-kind
200
THE PSYCHOLOGY OF GROUP PERCEPTION
category (Medin & Ortony, 1989). Psychological essentialism may then reflect peoples sense of deep ontological reality that is beyond human control and manipulation. Rothbart and Taylor (1992) suggested that there are two putative indices of psychological essentialism. One is the inductive potential among group members, and the other is the unalterability of group properties. It is note worthy that Haslam, Rothschild, and Ernst’s (2000) empirical study also found similarity and unalterability as underlying folk beliefs about social ontology. As other chapters of this volume will cover some details, I discuss them only briefly below. The inductive potential of a category is the extent to which the category is believed to act as a basis of inductive reasoning. So, for instance, if the observation of one instance of a category is believed to be generalizable to another instance of the category, the inductive potential of the category is said to be large. Arguably, one aspect of the inductive potential can be understood as a type of expected similarity between instances of the category. In fact, it is possible to argue that some significant aspects of Campbell’s conception of entitativity correspond to psychological essentialism as indexed by similarity and unalterability. First, Campbell’s move to use Gestalt principles to derive measures of entitativity amounted to operationally defining perceived entitativity in terms of, I would argue, different kinds of similarities. According to the principle of common fate, elements that belong to a putative entity move together in space and time. The similarity of elements was a second Gestalt principle, which is just that, the similarity of appearance. Proximity was another principle, which meant that elements of a putative entity tend to be found closer together in space. Each of them is a version of similarity: similarity in motion, appearance, and location. When translated to perceived group entitativity, these principles may be understood as the similarity in group members’ goal-directed activity, member characteristics such as appearance and disposition, and the physical location of the group members. It suffices to say that one strand of research on entitativity clearly uses similarity as a way of measuring (e.g., Brewer, Weber, & Carini, 1995; McGarty, Haslam, Hutchinson, & Grace, 1995) and manipulating (e.g., Dasgupta, Banaji, & Abelson, 1999) perceived group entitativity. Similarity is also theorized to mediate the process of group impression formation as opposed to individual impression formation (e.g., Hamilton & Sherman, 1996; McConnell, Sherman, & Hamilton, 1994b; McConnell, Sherman, & Hamilton, 1997; Susskind, Maurer, Thakkar, Hamilton, & Sherman, 1999). A second aspect of psychological essentialism of a social category can be regarded as the perceived unalterability of group properties. This interpretation of psychological essentialism of a social category turns on a certain reading of Rothbart and Taylor (1992). In their writing Rothbart and Taylor were talking more about whether a member of a social category can move into another social category, that is, the permeability of group boundaries. However, it is possible to conceptualize permeability as the extent to which group properties can be altered. If it is possible to alter properties that are characteristic of a group into those properties that are characteristic of another group, a member of the former group can become a member of the latter group. It is just that some social categories such as gender and racial categories may be seen to have some unalterable properties that make it impossible, say, for a man to become a woman or a European to become an Asian. In other words, some social categories may be seen to belong in the same realm of deep ontological reality as do natural kinds, with unalterable properties that may be inherited or otherwise so entrenched that they are almost impossible to change. It is in this sense that the Rothbart-Taylor formulation can be interpreted as suggesting that an aspect of group entitativity as the real existence of a group is indexed by the perception of unalterability of the properties of the group. Some researchers highlighted this aspect of group entitativity in their research (e.g., Levy & Dweck, 1999; Levy, Stroessner, & Dweck, 1998).
CULTURE, COMMUNICATION, AND ENTITATIVITY
201
Entitativity as Agency At this point, it is instructive to contrast the present conception of entitativity with that emphasized by Hamilton, Sherman, and their colleagues (e.g., Hamilton & Sherman, 1996; Lickel, Hamilton, Wieczorkowska, Lewis, Sherman, & Uhles, 2000). According to their conception, group entitativity is determined by the extent to which the group has an internal structure that makes group members interdependent of each other. The hallmark of an entitative group in this view would be a shared group goal, role differentiation, and coordination and interdependence among group members’ activities. To put it differently, their argument has been that when a group is seen to be a social agent, it is seen to be entitative. As Brewer (this volume) noted, entitativity as essentialism and entitativity as agency are two overlapping but distinct conceptions of entitativity. Indeed, Campbell’s original conception did imply that agency may indicate entitativity. Recall Campbell’s speculation that common fate may index an entitative group. This can be interpreted as his suggestion that, if members of a group exhibit similar goal-directed activities and attain a common outcome through them, the group may be entitative. In other words, if members of a group are seen to be thriving for a shared group goal, it may be seen to be entitative. It is interesting to speculate on the relationship between entitativity as essentialism and entitativity as agency at this point. First, the presence of agency may sometimes imply the presence of an essence. That is, it is possible that an agentic group is perceived to have an underlying essence. The shared goal in an agentic group may be seen to stem from a shared underlying essence, and hence the agentic group may be attributed an underlying essence. However, the presence of an essence may not always imply agency. Note that most physical objects such as a diamond are seen as natural kinds and likely seen to have essences, but agency is unlikely to be attributed to them. In other words, a number of natural-kind categories that may be understood in essentialist terms may not be seen to have agency. So, too, some social categories. Some social categories may be attributed agency, which then may be seen to possess essences. However, some social categories may be seen to possess essences but not agency. Do Culture and Communication Affect the Perceived Entitativity of Social Categories? Viewed in the current perspective, there emerges a theoretical possibility regarding relationships among culture, communication, and perceived entitativity of social categories. Of the three links portrayed in Figure 14.1, I will concentrate on the links involving social reality and entitativity because of the theoretical focus of this chapter. Still, I may be forgiven in touching on the culture-communication link briefly here. Kashima (2000b; also see Lyons & Kashima, 2001) has argued that interpersonal communications tend to reproduce information that is generally in line with shared cultures. It was shown that when people in a communication chain transmit a story containing information that is consistent and inconsistent with shared cultural stereotypes about genders, stereotype consistent information tends to be retained in the story in the long run. In other words, information circulated in society may more likely confirm, rather than challenge, their shared culture and social reality. This research is currently ongoing. Interested readers are referred to a review by McIntyre, Lyons, Clark, and Kashima (2004). Let us now turn to the two other links within the triad that have received some empirical attention, that is, the communication-entitativity and culture-entitativity links. Communication and Perceived Entitativity Does communicating about a social category increase the communicator’s perception of the entitativity of the group? In the first section it was suggested that communicating about something can increase the
202
THE PSYCHOLOGY OF GROUP PERCEPTION
psychological sense of its being real. By the same reasoning it is possible to postulate that communicating about a social category can increase the perceived entitativity of this category. Circumstantial evidence suggests that this may be the case. According to the cognitive tuning literature (e.g., Zajonc, 1960; see Guerin & Inness, 1989, for a review), the act of communicating about a person polarizes the person impression and leads to a stronger dispositional and causal inference about the target person. Likewise, the act of communication may polarize the group impression and lead to a stronger dispositional and causal inference. As Park and Judd (1990) showed, polarized group impressions are highly related (or perhaps identical) to the extent to which the characteristic attributed to the group is prevalent among group members. My own research (Kashima & Vears, 2000) has also found that the polarized group impression and perceived prevalence (percentage of people in the group that possess a given characteristic) are closely related to the perceived generalizability (likelihood that the observation of a group member can be generalized to another group member). Given these, it is likely that communication about a group can increase the perceived similarity among group members. Thompson, Judd, and Park (2000) reported that this is indeed the case. Although their research is framed in terms of stereotype formation through communication, in two experiments they examined the effects of communication on the perception of similarity among group members, one aspect of perceived entitativity as I discussed it before. In Experiment 1, nine-person groups were formed with a first set of three receiving behavioral information about a group, but a second set of three receiving the first three participants’ written communications about the target group, and likewise a third set receiving the communication of the second set of three. One half of the groups received no other information, whereas the remainder received the communications as well as the original behavioral information. Group impressions reported by the second and third sets of participants were more polarized and their perception of the prevalence of the group characteristics was greater than those of the first set of participants. It is interesting to note that this happened regardless of whether additional behavioral information was provided or not. Although this experiment showed that group impressions formed on the basis of communicated information tend to be polarized and homogeneous, it does not clearly show the effect of communication on the communicators impressions. This latter question was addressed in their second experiment. They compared the conditions in which people formed their group impressions with or without group discussions and showed that group impressions were more polarized and homogeneous when they had group discussions than when they did not. Extending the Thompson et al. research in two ways, Kashima and Vears (2000) showed that communication tends to essentialize group impressions. In two experiments, participants were told either to memorize or to communicate about a group when they received behavioral information. In the communication condition, they did not have a face-to-face communication with another person but wrote short letters to their friends and relatives. Group impressions were found to be more polarized and homogeneous in the communication condition than in the memory condition, showing that writing communications alone is sufficient to produce the effect. More importantly, it was shown that participants in the communication condition believed more strongly that the specific target group’s characteristics cannot be changed than in the memory condition. The perception of unalterability of the characteristics of the target group was measured by items modified from those originally designed to tap people’s implicit theory about the unalterability of dispositional characteristics such as personality traits and intelligence (Dweck, 1999; modified from Chiu, Hong, & Dweck, 1997). This latter finding is particularly significant as it shows that the effect of communication extends beyond the perceived similarity and includes the second aspect of psychological essentialism, namely, perceived unalterability.
CULTURE, COMMUNICATION, AND ENTITATIVITY
203
Culture and Perceived Entitativity Does culture affect the perceived entitativity of groups? Given the cross-cultural difference in individualism and collectivism (e.g., Hofstede, 1980; Triandis, 1995), it is possible that people from different cultural backgrounds may perceive different degrees of entitativity in different types of social aggregates. In addition, if communication about groups universally increased their perceived entitativity, social categories may be seen to have greater entitativity in East Asia than in Western European—based cultures. This is because social identities (i.e., one’s group memberships) are often communicated to others in East Asian cultures, but this communicative practice is not as prevalent in North America and Western European countries (for a review, see Triandis, 1995; Kashima, 2001). At this juncture, Lickel et al.’s (2000) research is significant. They showed that both North Americans and Poles (often regarded as less individualistic than North Americans) made similar distinctions among different types of groups (e.g., task group, social category). Although their study does not address directly cultural differences or similarities in the level of perceived entitativity and it is premature to conclude the same typology of social groups exists in other parts of the world, their results suggest that the typology of social groups may be similar across cultures and that this provides a basis for cross-cultural comparisons in the perceptions of groups. Then, is it the case that groups are perceived to be more entitative in collectivist cultures than in individualist cultures? There are two reasons for suspecting that this may be the case. First, collectivist cultures are theorized to emphasize the importance of social groups (Hofstede, 1980; Triandis, 1995). This may mean that people in collectivist cultures regard social groups as part of their social reality. Another reason for this postulate is the effect of communication on perceived entitativity. It has been shown that people in collectivist East and South Asian cultures tend to describe themselves by using social identities, namely, their group memberships. For instance, Trafimow, Triandis, and Goto (1991), Rhee, Uleman, Lee, and Roman (1995), and Dhawan, Roseman, Naidu, Thapa, and Rettek (1995) reported respectively that Chinese, Korean, and Indian students were more likely to describe themselves by using social identities and social categories than their North American counterparts. These findings can be interpreted as suggesting that East and South Asians are more likely to talk about those groups of which they are members. If communicating about a group tends to essentialize the communicator’s group impressions in all cultures, it may indeed be the case that East and South Asians may have higher degrees of perceived entitativity for social groups, especially, for those social groups about which they talk. The existing literature on cross-cultural group perceptions suggests that this may be the case at least for Chinese culture. Menon, Morris, Chiu, and Hong (1999), in their Study 1, examined a U.S. and a Japanese newspaper’s treatment of rogue trader scandals (two Japanese and two Western cases). In all cases, the Japanese newspaper referred to the organization for which the trader worked more often than the trader, but the U.S. newspaper referred to the trader more often than to the organization. In Study 2, Hong Kong and U.S. students responded to the description of a case in which a group members behaviors caused negative consequences for the group. Hong Kong students tended to attribute more to the dispositions of the group and less to the dispositions of the individual than their American counterparts. Study 3 examined Hong Kong and U.S. students’ reactions to three cases in which either an individual’s or a group’s action caused negative consequences. The Americans’ dispositional attributions were greater for the individual than for the group, whereas the Hong Kong students’ dispositional attributions were similar for the individual and for the group (also see Chiu, Morris, Hong, & Menon, 2000). If the degree of dispositional attributions can be used as a measure of entitativity, the evidence suggests that the individual is clearly seen to be more entitative than a group in North America, but the individual-group difference is much less pronounced in Hong Kong. In a recent cross-cultural study, a multicultural team of researchers (Kashima, Kashima, Chiu, Farsides, Gelfand, Hong, et al., 2001) more directly examined the culture-entitativity link. Perceived entitativity of
204
THE PSYCHOLOGY OF GROUP PERCEPTION
four targets (individual, family, friendship group, and society) were examined in three English-speaking cultures (Australia, U.K., and U.S.) and three East Asian cultures (Hong Kong, Japan, and Korea), as well as two continental European cultures (Belgium and Germany). University students responded to three measures of perceived entitativity that tapped the perceptions of inductive potential (how likely one observation about a given target is generalizable to another observation), unalterability (whether properties of a given target can be changed), and agency (how likely a given target is seen to have agency). Perceived agency was examined in terms of the appropriateness of attributing mental states such as thinking, wanting, and intending to each of the targets. Factor analyses showed that the measures of perceived inductive potential, unalterability, and agency were largely independent of each other, and the four targets (individual, family, friendship group, and society) were compared on these three dimensions. With regard to perceived inductive potential and unalterability, the individual was universally seen to be more entitative than the groups (family, friends, and society). There was a minor cross-cultural variation in that the family was perceived to be as entitative as the individual in Japan and the English-speaking cultures on the dimension of inductive potential. What distinguished East Asian cultures and the five other Western cultures most clearly was the dimension of perceived agency. In the Western cultures the individual was seen to be more agentic than some social groups. In some cases intimacy groups such as the family and friendship groups were perceived to be as agentic as the individual; however, society was always seen to be less agentic than the individual. In contrast, in the East Asian cultures all targets were perceived to be equally agentic. In other words, East Asians did not differentiate the individual, small intimacy groups, and society by the extent of perceived agency. CONCLUDING REMARKS All in all, research on group entitativity and psychological essentialism puts the question of social ontology of human groups in particular, and the issue of social reality in general, squarely on the agenda of social psychological research. Campbell’s conception of entitativity, Rothbart and Taylor’s discussion of psychological essentialism, and the research on communication and social reality give us the necessary conceptual backing to conduct empirical research on the construction and maintenance of social reality within a social cognitive framework. What makes it worthwhile asserting this point at the beginning of the 21st century is the strain that the concept of “social reality” has suffered in social psychology in the past few decades. In the last quarter of the 20th century social psychology, and the social sciences more generally, witnessed an intellectual challenge to its theoretical and epistemological fronts. This Zeitgeist appeared most clearly and symptomatically in Gergen’s (1973, 1985, 1994) challenge to the “scientific” experimental method of inquiry as a privileged method, and the social constructionist argument that social interaction can construct reality, even reality as claimed by scientists. I think that social constructionists are correct about one of their claims: Humans do construct social reality. However, to say so does not mean that we must endorse all of their claims. I for one do not endorse the extreme epistemological claim that experimentation does not permit an inquiry into social psychological processes. On the contrary, I believe that the construction of social reality can be examined using a variety of methods, including the method of experimentation. My central claims here have been that humans construct social reality through the use of culture and symbolic capacity mainly in communication. We maintain this constructed social reality in part by generating and sustaining the psychological sense of social reality. Although much research still needs to be conducted, there is some suggestive evidence for these claims in the area of communication and perceived
CULTURE, COMMUNICATION, AND ENTITATIVITY
205
entitativity as well as in the research on culture and group perceptions. In the end, the social construction of social reality can, and should also, be studied within the social cognitive research tradition.
15 Group Socialization, Uncertainty Reduction, and the Development of New Members' Perceptions of Group Variability CAREY S.RYAN University of Nebraska at Omaha DEBBIE R.ROBINSON LESLIE R.M.HAUSMANN University of Pittsburgh
The process by which people become full members of existing social groups has long been of interest to organizational researchers. The socialization process, including its content (Chao, O’Leary-Kelly, Wolf, Klein, & Gardner, 1994), harshness (Irving & Meyer, 1994), and uncertainty (Louis, 1980), influences newcomers’ attitudes toward the group as well as their motivation and ability to achieve group goals (Saks, 1995). At the same time, stereotyping researchers have long been interested in how people come to perceive ingroups differently from outgroups. In particular, a great deal of research during the last two decades indicates that an important determinant of perceived group variability is simply whether the perceiver belongs to the group being judged; ingroups are often perceived to be more variable than outgroups (Park & Rothbart, 1982). Our goal in this chapter is to consider both the organizational and stereotyping literatures in an effort to understand how the process of socialization influences perceptions of ingroup and outgroup variability. We begin by reviewing previous research concerning the factors that influence perceived group variability. We then describe the socialization process and the uncertainty that new members of existing groups experience after joining a group. Socialization and uncertainty are, of course, multifaceted phenomena, and any description of them is thus likely to be incomplete. Our goal is to identify those aspects of the socialization process that seem most likely to influence perceptions of ingroup and outgroup variability. We then attempt to integrate the relevant findings from these two literatures into a group socialization account of stereotype development. Finally, we describe the results of two studies that provide some empirical support for our theoretical analysis, including a detailed longitudinal study of new members’ perceptions of actual social groups and a study in which three key variables were experimentally manipulated.
PERCEPTIONS OF GROUP VARIABILITY
207
CAUSES OF PERCEIVED GROUP VARIABILITY Researchers have sought to understand the factors that influence perceived group variability by focusing primarily on the cognitive processes that underlie variability judgments. Three general issues are particularly relevant to this chapter: the type of information that people encounter, the manner in which the information is cognitively organized, and the perceiver’s status as an ingroup or outgroup member. Type of Information Encountered Variability perceptions differ not only as a function of the actual variability of group members (Park & Hastie, 1987; Park, Judd, & Ryan, 1991), but also as a function of the form and sequence in which grouprelevant information is encountered. In particular, it appears that people who initially learn about a group from information about the group as a whole (i.e., abstraction-based information) as opposed to information about individual members of the group (i.e., instance-based information) perceive the group to be less variable (Park & Hastie, 1987). This appears to occur even when perceivers ultimately encounter identical information about the group. The implications of this work for understanding ingroup-outgroup differences in perceived variability are significant. People who learn about an ingroup are more likely to do so from individual group members initially, whereas people who learn about an outgroup are more likely to do so from group-level information initially, for example, portrayals of the group provided by the media or by other ingroup members. This research thus suggests that ingroups are likely to be perceived as more variable than outgroups. Other characteristics of the information encountered are also likely to be important. For example, perceivers may process stereotype-incongruent information more carefully than stereotype-congruent information, resulting in greater perceived variability (Garcia-Marques & Mackie, 1999). And perceivers may focus on information that is central (vs. peripheral) to the group’s identity in an effort to form an impression of the group as a whole, leading them to perceive less variability (cf. R.Brown & WoottonMillward, 1993). As with group-and individual-level information, such effects are likely to depend on the point in the learning process at which they are encountered. But perceptions of group variability do not depend only on the type of information encountered and the point at which it is encountered; they also depend on how the information is cognitively processed and stored in memory. Knowledge Structures A great deal of research indicates that people develop more differentiated cognitive structures to represent information about ingroups than outgroups and that more differentiated cognitive structures lead to greater perceived group variability. However, our understanding of these processes remains limited. Differentiation has been conceptualized (and operationalized) in various ways. Some research has focused primarily on the hierarchical nature of knowledge structures, for example, the extent to which the overall group is represented in the form of subgroups (Park, Ryan, & Judd, 1992) or specific exemplars (Linville, Fischer, & Salovey, 1989). Other research seems to focus on how the information at a given level in the hierarchy is organized in memory. For example, greater complexity of one’s knowledge structure may be indicated by descriptions of a group that utilize attribute dimensions that are relatively uncorrelated (Linville, Fischer, & Yoon, 1996). These approaches are, of course, not mutually exclusive. And as we indicated above, our understanding of these issues remains limited. The important points for now are that much of the perceived group variability research has focused on the types of cognitive processes and structures that underlie variability perceptions. Although we do not know for sure how people represent variability information or
208
THE PSYCHOLOGY OF GROUP PERCEPTION
the precise ways in which a cognitive structure may be considered differentiated, people who develop more differentiated cognitive structures, as they have thus far been defined, perceive greater variability. Importantly, developing more differentiated cognitive structures and perceiving greater variability do not automatically result from greater familiarity with the group. Longitudinal studies have demonstrated that, as new group members’ length of involvement and familiarity with their group increase over time, perceived variability of the group decreases (R.Brown & Wootton-Millward, 1993; Oakes, Haslam, Morrison, & Grace, 1995; Ryan & Bogart, 1997; but see Linville et al., 1989). Furthermore, people who have identical information about individual members of a group have been shown to perceive the group as more variable when they (i.e., the perceivers) are instructed to organize the information into meaningful subgroups than when they are not (Park et al., 1992). Thus, the ways that people process and organize the information they encounter, rather than familiarity per se, appear to be the more important determinants of perceived variability. An obvious question, of course, is, What causes perceivers to develop more differentiated cognitive structures to represent a group? Clearly, perceivers must be motivated to organize group-relevant information in ways that allow that information to be utilized more effectively (at least from the perceivers perspective) in the future. Otherwise, there would be little reason to expend cognitive resources in an effort to organize information into complex and meaningful structures. One reason perceivers may do so is that the information concerns a group that is important to them. Group Membership Perceivers may deem a group to be important for a variety of reasons. For example, employers may perceive college graduates to be an important source of new members and television viewers may perceive broadcast journalists as an important source of information. Perceivers may also consider a group to be important simply because they belong to it; that is, the group is an ingroup. Presumably, perceivers are motivated to make fine distinctions among ingroup members because such information is useful in daily interactions. This is less likely to be the case for outgroups inasmuch as daily activities are less likely to involve encounters with outgroup members. According to perceived group variability researchers (e.g., Park, Judd, & Ryan, 1991), this difference in motivation leads perceivers to develop more differentiated cognitive structures to represent information about the ingroup than the outgroup, resulting in outgroup homogeneity. Note that the conceptualization of group membership as ingroup versus outgroup is extremely simple. It does not take into account intragroup factors, such as the social norms that may be operating among ingroup members, whether the perceiver is a brand-new member of the group or an old-timer, role demands, or, as we suggested above, the fact that perceivers may consider a group— even an outgroup—to be important because of its ability to provide desired resources (e.g., new members or information). To be sure, some research has examined the degree to which perceivers identify with their ingroup (Tajfel & Turner, 1986) or adopt a particular social identity through self-categorization (Turner, Hogg, Oakes, Reicher, & Wetherell, 1987) as opposed to simple ingroup versus outgroup membership. However, unlike the perspective we present in this chapter, this research has largely focused on understanding the effects of the intergroup context on intergroup perceptions (Doosje, Spears, Ellemers, & Koomen, 1999; but see Castano & Yzerbyt, 1998). The simple conceptualization of group membership that has been used in most research may prevent full understanding of the outgroup homogeneity effect and perceived group variability more generally. Indeed, as researchers have begun to examine a wider range of groups and intergroup contexts, exceptions to the outgroup homogeneity effect have emerged (Doojse et al., 1999). For example, when perceivers are
PERCEPTIONS OF GROUP VARIABILITY
209
members of a numerical minority (Simon & Brown, 1987), belong to a group whose purpose is to understand what outgroup members are like (Ryan, Robinson, & Hausmann, 2001), or are very new members of the group (Ryan & Bogart, 1997), they appear to perceive greater variability among outgroup than among ingroup members. The outgroup homogeneity effect also appears not to occur among AfricanAmerican and white American college students (Judd, Park, Ryan, Brauer, & Kraus, 1995; Ryan, Judd, & Park, 1996). Indeed, this work indicates that African Americans’ and White Americans’ intergroup perceptions may be influenced by socialization processes that emphasize different ethnic perspectives, with a multicultural perspective more characteristic of African Americans’ experiences and a colorblind perspective more characteristic of White Americans’ experiences. It seems, then, that perceived group variability may be influenced by intragroup experiences as well as by ingroup versus outgroup membership. But simply expanding the types of groups that are examined or the range of intergroup contexts in which judgments are made can never fully capture the complex nature of what it means to belong to a group. Neither can assessing the degree to which perceivers identify with their ingroup or categorize themselves as group members. Other group members are fundamentally involved in such processes. For example, the adoption of a given social identity, not to mention social acceptance, generally requires affirmation from others that one possesses the criteria for group membership. Such affirmation does not typically occur immediately upon entry into a group. What is needed is a model of group membership and belonging in which intragroup experiences and social psychological processes that unfold over time assume more prominent roles. Group socialization theory provides such a model. GROUP SOCIALIZATION An influential theory of group socialization comes from Moreland and Levine (1982; 1989; Levine & Moreland, 1994), who argue for five phases of group membership: investigation, socialization, maintenance, resocialization, and remembrance. The theory posits that as individuals move in and out of each phase, their relationships to the group and role demands change, and different issues become more or less important. Although there is virtually no empirical work examining the issue, we believe that perceptions of ingroups and outgroups are likely to change during the first two and last two phases, with the greatest change occurring during the investigation and socialization phases. Our present focus concerns the socialization phase. The socialization phase begins when individuals enter the group. During this phase, both the group and the individual member try to change one another so that their relationship becomes more rewarding (Moreland & Levine, 1982; 1989). The group wants the new member to help achieve its goals, while the new member wants the group to help satisfy his or her needs. To the extent that both the group and the new member are satisfied with the contributions of the other, their levels of commitment to each other increase. When the commitment of each reaches sufficient levels, acceptance occurs, and the new member becomes a full member of the group. An essential task of group members during the socialization phase is informaton acquisition (V.D.Miller & Jablin, 1991). Indeed, the length of the socialization phase varies considerably (Moreland &: Levine, 1982), at least partly as a function of the level of knowledge and sophistication of behaviors that are required for membership. In some groups, such as sororities and fraternities and entry-level work groups, the socialization phase may last only a few months. For other groups, for example, those entering a profession such as law, university teaching, or medicine, the socialization process may last several years. During this socialization phase, new members and senior members engage in behaviors that not only influence new members’ abilities to contribute to the group (Moreland & Levine, 1982), but also the
210
THE PSYCHOLOGY OF GROUP PERCEPTION
impressions of ingroups and relevant outgroups that new members develop (Levine, Moreland, & Ryan, 1997; Ryan & Bogart, 1997; 2001). The socialization phase is also characterized by relatively negative experiences. New members initially tend to have unrealistically positive expectations about their future experiences in groups (Wanous, 1977; Wanous, Poland, Premack, & Davis, 1992). New members may also have socialization experiences that highlight their lower status and lack of true acceptance. Some experiences, for example, being asked to perform menial tasks or even the violation of psychological contracts (Robinson & Rousseau, 1994), are not necessarily intended to emphasize status differences. In some cases, however, such as hazing, senior group members’ intentions seem quite clear. New members must demonstrate their commitment to the group before full acceptance is granted. Interestingly, although the group’s intention is to increase commitment, cohesion, and thus positive attitudes toward the group, negative socialization experiences are more likely to result in negative attitudes (Irving & Meyer, 1994; Lodewijkx & Syroit, 1997; Robinson & Rousseau, 1994).1 As our analysis thus far implies, the socialization phase is also marked by uncertainty. New members are uncertain of what the group and its members are really like and whether the group will accept them. And it is actually the desire to reduce uncertainty that we believe plays the critical role in the development of group variability perceptions. UNCERTAINTY REDUCTION Numerous researchers and theorists have described the high levels of uncertainty that characterize the socialization phase of group membership (Falcione & Wilson, 1988; Lester, 1987; Louis, 1980; Moreland & Levine, 1982; 1989; Saks & Ashforth, 1997). New members have a great deal to learn about the group. Obviously, they must learn the appropriate skills, behaviors, and procedures associated with their roles in the group. Often, but certainly not always (Gruenfield & Fan, 1999), such information is explicit, and there are formal mechanisms in place to train new members. Less obvious, however, is the need for new members to learn about the group. New members must learn how other members are similar to and different from one another, how the group as a whole is similar to and different from other groups, what roles individual members (including themselves) play in the group, and so forth. This type of information is equally important but more difficult to learn, because it is usually much less explicit. New members are also likely to be uncertain about whether they will ultimately be fully accepted by the group (Moreland & Levine, 1989), which we believe to be a critical motivating factor. New members typically do not have full membership status, which involves both formal acceptance (e.g., the successful completion of a probationary period) and informal acceptance—the feeling of truly being one of the guys or gals. Indeed, Kerr (1999, p. 112) has referred to social exclusion or rejection as the most substantial social sanction—the “social death penalty.” Social exclusion—being excluded from a group that one wants to join —is psychologically painful (K.D.Williams & Zadro, 2001). We believe it is this desire to be accepted that motivates new members to go to a great deal of effort to learn about and understand the group. In short, new members experience uncertainty, because they wish to be accepted, but they are not sure what the group is really like and what is required of them as group members. They are therefore uncertain about what is required for acceptance to occur. According to uncertainty reduction theorists (Falcione & Wilson, 1988; Lester, 1987; Saks & Ashforth, 1997), new members, like others (Sorrentino & Roney, 2000; Sorrentino & Short, 1986), are motivated to reduce their uncertainty so that the environment is more predictable, understandable, and controllable. New members can reduce uncertainty about the environment by acquiring information about the group, including
PERCEPTIONS OF GROUP VARIABILITY
211
its characteristics, norms, expectations for new members, and so forth. Perhaps the most important source of information is direct interaction with other group members (Miller & Jablin, 1991; Saks & Ashforth, 1997). Direct interaction can provide information concerning the group, such as norms and role demands. It can also provide clues as to whether new members’ behaviors are moving them toward (or away from) acceptance by the group. As new members acquire such information, their uncertainty decreases, they become more adept in their roles, and they are more likely to remain in the group (Mignerey, Rubin, & Gorden, 1995; Miller & Jablin, 1991; Morrison, 1993).2 We have now laid the groundwork for our group socialization account of stereotype development, which focuses on the development of perceived group variability and ingroup bias. Thus far we have discussed each area of work in isolation. We now turn to a discussion of our perspective in which this work is integrated more explicitly. DEVELOPMENT OF NEW MEMBERS’ PERCEPTIONS OF GROUP VARIABILITY: A GROUP SOCIALIZATION PERSPECTIVE Individuals who enter existing groups want ultimately to be fully accepted by the group. But they are uncertain about whether acceptance will occur and what is required to “fit in.” New members are therefore especially motivated to learn what the group is like and what is required to gain acceptance. At the same time, individuals at this stage may have little information on which to base their impressions and expectations. Even those who have a great deal of information are likely to be motivated to acquire more. Information acquired prior to entry is generally incomplete and often positively biased (Wanous, 1977). Upon entry into the group, then, new members typically have relatively vague notions of what the group and its members are like. One way to deal with this state of affairs is simply to gather relevant information so that one’s uncertainty about what the group is like is reduced and one’s understanding of what is required for acceptance is increased. Such information may include information about the mundane, for example, the procedures for ordering supplies. But it is also likely to include information about the characteristics of the group and its members, as well as information about how new members fit in and what is required for full acceptance. This latter type of tacit information takes time to acquire inasmuch as it must usually be extracted from, for example, orientation and formal training; written documents concerning a group’s history, purpose, and procedures; observation of other group members’ behaviors; and direct interactions with group members and with other individuals who are related to the group. In any case, one result of this learning process is the development over time of cognitive structures that represent information about the group. How might the specific process of forming such structures occur? We believe that new members begin by identifying common themes or general similarities among group members. For example, new members may feel it more important to know whether the group as a whole endorses liberal policies and, if so, which specific policies, as opposed to whether a particular group member is more or less liberal than another. In other words, new members focus on developing their impressions of the group by first developing an organized, relatively elaborate, and meaningful impression at the superordinate level. Such a structure allows new members to utilize the information more effectively in the future and to assimilate the vast amount of new information that they will continue to encounter. Focusing on group characteristics and similarities among group members is likely to result in lower perceived group variability (Park et al., 1992; Worchel, Coutant-Sassic, & Grossman, 1992) and increased confidence in one’s understanding of the group and its members (Ryan, Judd, & Park, 1996; Ryan et al., 2001) as socialization proceeds (Ryan & Bogart, 1997). Note that this does not mean that new members
212
THE PSYCHOLOGY OF GROUP PERCEPTION
develop less accurate group impressions (Ryan, 2002). Indeed, we expect that as socialization proceeds new members develop more refined, more elaborate, and more accurate group impressions, especially for the ingroup—even though they perceive the groups, on average, to be less variable and less positive over time. Of course, understanding what individual group members are like is also important to new group members. We believe, however, that what is of greater importance initially are the sorts of characteristics that group members have in common, how they contribute to the group, and so forth. This primary focus on the group as a whole seems likely to shift only gradually over time toward understanding the particular ways in which individual group members differ. We suspect that such a shift begins relatively late in the socialization phase, perhaps not becoming the primary focus until sometime beyond the socialization phase. This process, that is, the elaboration of group-level impressions during socialization and gradual shift over time toward individual differences among group members, seems more likely to occur for ingroups than for outgroups, because information about the ingroup plays a more important role in daily social interactions. It is these daily social interactions that ultimately determine whether new members are fully accepted by the group. New members may thus perceive both their ingroups and outgroups to be less variable later (rather than earlier) in the socialization phase, because their primary focus during the socialization phase is on understanding the groups overall. But the reduction over time is expected to be greater for the outgroup than the ingroup. We suspect that at some point beyond the socialization phase perceptions of ingroup variability may increase as a result of the shift in focus to differences among individual group members. In sum, uncertainty about being accepted into a new group and a concomitant lack of knowledge about the group lead new members to focus on understanding the group as a whole, resulting in decreasing perceived variability during socialization. This reduction is greater for the outgroup, because less elaborate group-level impressions may be adequate for the outgroup, whereas a more elaborate group-level impression is needed for the ingroup. The process of developing group-level structures enables new members to organize the information they encounter throughout the socialization period and beyond. Indeed, because such structures can facilitate the organization and assimilation of new information, we suspect that these elaborate group-level impressions may ultimately allow the diversity within the ingroup to become apparent (cf. Attneave, 1957). In other words, focusing on the group as a whole may ultimately (i.e., at a later in time beyond the socialization phase) lead to greater perceived variability when that focus is part of an effort to develop an organized and refined impression of the group. In general, we would also expect new members to judge the ingroup less positively as the socialization phase proceeds. Recall that, in addition to uncertainty about the group and whether they will be accepted, new members typically have unrealistically positive expectations about their future experiences in the group (Wanous et al., 1992) and may have relatively negative experiences, for example, experiences that highlight their lower status in the group (Irving & Meyer, 1994). Such experiences, coupled with increased knowledge of the more negative characteristics of the group, are likely to result in less positive perceptions of the ingroup and thus, decreased ingroup bias, as the socialization phase proceeds. Of course, groups may attempt to provide more positive socialization experiences and, to the extent that they are successful, attenuate this general pattern (Irving & Meyer, 1994). LONGITUDINAL STUDY OF ACTUAL SOCIAL GROUPS The group socialization perspective that we have developed was actually stimulated by findings from a longitudinal study of new group members’ ingroup and outgroup stereotypes (Ryan & Bogart, 1997). We asked the new members (i.e., pledges) of four campus sororities to judge their ingroup and three outgroup sororities four times during the year. Participants completed several tasks that provided measures of
PERCEPTIONS OF GROUP VARIABILITY
213
perceived group dispersion, positivity, and stereotypicality. Perceived group dispersion is what we generally think of as variability; it refers to perceptions of the variation of group members around the mean of the group. Perceived group positivity refers to how positive (vs. negative) a group is perceived to be. Perceived group stereotypicality refers to perceptions of the group’s central tendency on stereotypic versus counterstereotypic attributes. Stereotypicality has sometimes been referred to as a measure of perceived variability (e.g., Park & Judd, 1990). We believe that it is more appropriately characterized as a sophisticated measure of perceived central tendency, that is, one that controls for differences in the use of the response scale by including items that are both stereotypic and counterstereotypic of the group. In any case, it is clearly an important component of stereotypes that can be virtually uncorrelated with perceived group dispersion (e.g., Park & Judd, 1990). Moreover, results concerning this stereotype component are often different from those for dispersion so that research designed to understand change in perceived group variability, and stereotype development more generally, seem likely to be most informed by research that examines both. Indeed, in our longitudinal study of new sorority members, we obtained different results for dispersion and stereotypicality. Participants perceived their outgroups more stereotypically than their ingroups at every wave—the same outgroup homogeneity effect that Park and Rothbart (1982) found. And although perceived stereotypicality generally decreased during the socialization phase, there was no change in the ingroup— outgroup difference. Thus, the ingroup— outgroup difference in perceived stereotypicality was established early and remained relatively stable as socialization proceeded. In contrast, outgroups were perceived to be more dispersed than ingroups in the first wave of data—an ingroup homogeneity effect. And as socialization proceeded, participants judged their ingroups and outgroups, but especially their outgroups, to be increasingly less dispersed. Thus, these new members increasingly came to perceive both groups as less dispersed, but because this effect was much stronger for the outgroup, participants’ judgments moved in the direction of outgroup homogeneity. Indeed, by wave four, these new members had generally come to perceive their outgroups to be somewhat less dispersed than their ingroups, although the difference was not statistically significant. This decrease in perceived dispersion occurred at the same time that participants were becoming increasingly familiar with the groups, especially the ingroup. It is also consistent with the results of two other longitudinal studies of new members’ perceptions of ingroup variability (R.Brown & Wootton-Millward, 1993; Oakes et al., 1995). Finally, participants also exhibited ingroup bias at each wave but judged the groups to be increasingly less positive as socialization proceeded. The reduction in perceived positivity was especially true for judgments of the ingroup, which resulted in decreased ingroup bias as socialization proceeded. Participants’ self-ratings further indicated that they perceived themselves less positively over time on group-relevant attributes. We have since done further analyses of the same new members’ judgments to examine longitudinal changes in stereotype accuracy (Ryan & Bogart, 2001). As part of the same study, we asked the full members of each sorority to provide self-ratings on each of the stereotype-relevant attribute dimensions at the same four times that new members’ stereotypes were assessed. We used these self-ratings to compute the actual self-rated stereotypicality and the actual self-rated dispersion of each group at each wave. New members’ judgments of the groups were compared with these actual values, using discrepancy scores and within-subject sensitivity correlations (Ryan, Park, & Judd, 1996). According to our group socialization explanation, new members develop and refine their impressions of groups, especially ingroups, as socialization proceeds. Their goal is to understand what the groups, especially the ingroup, are like and how they can best fit into the ingroup and achieve acceptance. We might therefore expect new members to judge the groups, especially the ingroup, more accurately as socialization
214
THE PSYCHOLOGY OF GROUP PERCEPTION
proceeds. In other words, we expected to find evidence of increased accuracy even though these new members had judged the groups to be increasingly less variable and less positive as socialization proceeded (Ryan & Bogart, 1997). The data (Ryan & Bogart, 2001) did, in fact, yield evidence of increased accuracy over time, especially in judgments of the ingroup. Ingroup stereotypicality was judged more accurately than was outgroup stereotypicality on both discrepancy and sensitivity correlation measures, and although this ingroupoutgroup difference did not change, stereotypicality judgments generally became more accurate over time. These new members also underestimated ingroup more than outgroup dispersion at wave one, shortly after they had pledged the sororities. But ingroup judgments became more accurate, whereas outgroup judgments became less accurate during the socialization phase. The sensitivity correlation measures of dispersion accuracy further revealed increased accuracy as the socialization phase proceeded, although this change was equally true for the ingroup and outgroup. In sum, the data from this longitudinal study of real-world social groups seem best explained by a group socialization perspective. Both ingroups and outgroups were perceived to be less dispersed (especially the outgroup) and less positive (especially the ingroup) over time. Nevertheless, their perceptions of the groups, especially the ingroup, became more accurate. This increased accuracy over time suggests that new members were refining and elaborating their impressions of the groups at the same time that they perceived the groups to be increasingly less dispersed and less positive. EXPERIMENTAL RESEARCH We have also tested parts of our group socialization explanation in a laboratory experiment. We have suggested that new members want to be fully accepted by the group and are uncertain about whether and how it may be achieved. One way of reducing this uncertainty and ultimately maximizing the chances of acceptance is to learn what the group is like. Such information helps new members determine the sorts of things (e.g., behaviors) they must do to be accepted by the group. In one experiment, then, we specifically examined the effects of acceptance certainty on new members’ preference for group- versus individuallevel information about ingroups and outgroups. We expected that uncertainty about being accepted by a group would cause new members to focus initially on learning about the group as a whole as op posed to learning about individual group members, especially for the ingroup. We also expected that individuals who first focused on the group as a whole would later perceive the group, especially the ingroup, to be more variable. (We did not examine whether focusing on the group vs. individuals would lead to lower perceived group variability initially, which we would also expect.) In this study, then, we manipulated new members’ certainty about being accepted by the ingroup to determine whether uncertain participants would initially prefer group-level to individual-level information about the ingroup as compared with the outgroup. The design of the study was thus a 2 (“Pitt Film Club” vs. “Panther Movie Group” subject group) X 2 (certain vs. uncertain acceptance) X 2 (group- vs. individual-level information) X 2 (ingroup vs. outgroup judgments) factorial design, with repeated measures on the last two factors. Introductory psychology students who thought they would be interested in joining a campus movie group were asked to participate in a study that examined how people form impressions of new groups and their members. Students were told that immediately after the study, they would have an opportunity to meet other students who already belonged to the club to view and discuss a movie and talk about membership requirements (which may have actually increased participants’ interest in information about individual members). In fact, however, the two campus movie groups to which participants were randomly assigned
PERCEPTIONS OF GROUP VARIABILITY
215
did not exist. In addition to research credit, each participant was given $5 at the end of the study to compensate them for this deception. When participants arrived at the lab, they were told that they would learn more about the movie club and one other somewhat different movie club through the clubs’ webpages. Each participant was seated at a computer displaying the introduction to the study. The introduction briefly described a film club and explained that the club was recruiting new members. Participants were told that there was a pledge period of two months during which they would be able to participate in most of the club’s activities, including viewing and discussing movies with the rest of the members. Participants in the certain condition were told that about 95% of pledges were invited to become full members. Participants in the uncertain condition were told that only about 55% of pledges were invited to become full members. After reading the introduction, participants were asked to click on either “member” (in the case of the ingroup) or “guest” (in the case of the outgroup). The computer saved this information in a file. This step also initiated a timing sequence to track the order in which each bit of information was selected and the time that participants spent considering each bit of information. Next, participants saw a main page containing 10 icons that were labeled “about the group” and 10 that were labeled “about a member.” (Order of information type was counterbalanced across participants.) The icons provided links to descriptions of either the group as a whole or an individual group member. The grouplevel descriptions were consistent with the individual behaviors but did not provide information about the variability of group members. Participants were free to choose any number of information bits, in any combination, until they felt satisfied with their knowledge about the group, but they were told to stop as soon as they felt they had enough information. Each time participants clicked on an icon, an information bit appeared on the monitor. For example, one group-level bit of information describing the Pitt Film Club was: The club is a great place to discuss movies. Overall, the members are pretty casual and friendly, and they’re generally a lot of fun to be with; they like to joke around. And one individual-level bit of information describing a member of the Panther Movie Group was: Dana is on the debate team. She’s a great person, strong and knowledgeable but very good-hearted. She’s the one who lets everyone know about special showings and upcoming movies. After considering a given bit of information, participants returned to the list of icons to select another group- or individual-level bit of information. Participants used this same procedure to learn about both the ingroup movie club and an outgroup movie club. Following the computer task for both groups, participants completed a questionnaire to assess their perceptions of the stereotypicality and dispersion of each group. (Order of target group judgments was counterbalanced.) Participants judged each group with respect to the same eight attributes. All attributes were positively valenced. But half were stereotypic of the Panther Movie Group and counterstereotypic of the Pitt Film Club (i.e., “into academics,” “ambitious,” “organized,” and “businesslike”), and half were counterstereotypic of the Panther Movie Group and stereotypic of the Pitt Film Club (i.e., “relaxed,” “informal,” “caring,” and “accepting of others”). Finally, participants completed a free recall task in which they were asked to list all of the information they could remember about the groups. The results of this study were generally supportive of our predictions. As Figure 15.1 indicates, participants selected more individual- than group-level information (Ms=5.08 and 4.66), F (1, 150)=4.18,
216
THE PSYCHOLOGY OF GROUP PERCEPTION
FIGURE 15.1. Mean number of information bits selected as a function of information type, ingroup versus outgroup, and certainty of being accepted by the group
p<.05. But this difference depended on acceptance uncertainty and ingroup versus outgroup, F (1, 150)= 4. 29, p<.05. Specifically, participants who were uncertain selected more group-level relative to individuallevel information for the ingroup than the outgroup, whereas the reverse was true for participants who were relatively certain of acceptance. Participants who were certain of acceptance selected more group-level relative to individual-level information about the outgroup than the ingroup. We also examined whether differences in the type of information participants focused on initially (i.e., the first six information bits that participants selected) predicted later judgments of group stereotypicality and dispersion (partialing out acceptance uncertainty and subject group). The latter judgments were provided after participants had seen and thought about all of the information they had selected—not just the first six information bits. What was most striking about these correlations was that participants who initially focused more on group-level information about the ingroup than the outgroup subsequently judged the ingroup to be less stereotypic (pr (153)=−.19, p<.05) and more dispersed (pr (153)=.23, p<.01) than the outgroup. Analyses of the recall data further indicated that participants who were uncertain had significantly better recall than did those who were certain and that recall was better for ingroup than for outgroup and for group-level than for individual-level information (all ps<.05). In sum, the level of certainty associated with joining a new group appears to affect the type of information that new members seek. Although participants selected more individual- than group-level information overall, their preference for group- relative to individual-level information was greater for the ingroup than the outgroup when they were uncertain about entry into the group. Indeed, when participants were certain of being accepted by the group, they selected somewhat more group-level information about the outgroup than the ingroup. Moreover, the type of information that participants preferred initially appeared to influence later judgments of group stereotypicality and dispersion. Participants who initially focused on group-level information later judged the ingroup to be more diverse and less stereotypic than the outgroup.
PERCEPTIONS OF GROUP VARIABILITY
217
CONCLUSION Our primary goal in this paper has been to outline a socialization explanation for the development of new members’ perceptions of group variability and positivity. We have suggested that the new members of existing social groups are uncertain about their environment, including what their new group and its members are like, and whether the group will accept them. This uncertainty and desire to be accepted leads new members to focus initially on the similarities among group members and on the development of meaningful group-level impressions of the ingroup and relevant outgroups. This focus results in lower perceived dispersion. Importantly, however, we also believe that new members refine and develop their impressions of the groups over time—even though they focus on group-level information and perceive less dispersion. It is just that they do so, at least in the early stages, by thinking about the group primarily at the group level. As group-level impressions develop, new members gradually shift their focus to developing a more detailed understanding of subgroups and individual group members, especially for the ingroup. We believe, then, that when perceivers focus on the group in an effort to develop an organized and refined impression of the group, it may ultimately, that is, at some later time, perhaps after the socialization phase, lead to greater perceived group dispersion. A more elaborate group-level impression seems likely to facilitate the processing and organization of new information. We presented longitudinal research that generally supports this perspective. A longitudinal study of new sorority members indicated that (a) new group members perceive outgroups and, to a lesser extent, ingroups, as increasingly less dispersed and less positive during the socialization phase (Ryan & Bogart, 1997), and (b) new members appear to refine their impressions of ingroups and, to a lesser extent, outgroups as socialization proceeds (Ryan & Bogart, 2001). In other words, new members’ perceptions of the groups, especially ingroups, became more accurate over time. We also presented an experimental study, indicating that new members who were uncertain about being accepted by the group initially selected more groupversus individual-level information about the ingroup more than the outgroup, whereas the reverse was true for new members who were relatively certain of acceptance. Those who initially focused on group-level information about the ingroup more than the outgroup were ultimately more likely to perceive the ingroup as more variable than the outgroup. Of course, we consider this area of work to be in its earliest stages. The process of socialization is complex, and we have examined only a small part of it. For example, we have not yet identified the ways in which group impressions may be become more elaborate over time at the group level, the ways that groups may transmit group knowledge and culture to new members, or the specific consequences of this transmission process for new members’ group percep tions. And our experimental examination of new members’ efforts to learn about their group has thus far been limited to the effects of acceptance certainty on the preference for group-level information about the ingroup. We also believe that new members are much more active, attempting to influence, for example, other group members’ impressions of and behavior toward them, than our analysis here may suggest. Finally, for many groups, it is not clear when socialization ends and a new phase of membership begins or at what point perceptions of group variability become relatively stable. We believe research that examines these issues further is important for both theoretical and practical reasons. People choose to join many of the groups to which they belong, for example, the military (Biernat, Crandall, Young, Kobrynowics, & Halpin, 1998), educational groups (Alwin, Cohen, & Newcomb, 1991; Guimond, 2000), occupational groups (e.g., R.Brown & Wootton-Millward, 1993; Chao et al., 1994; Irving & Meyer, 1994; Lee, 1994; Ryan et al., 2001), sports teams (Hoover, 1999), and sororities and fraternities (Park & Rothbart, 1982; Ryan & Bogart, 1997; 2001). These groups and their specific-practices clearly vary
218
THE PSYCHOLOGY OF GROUP PERCEPTION
in important ways. Yet membership in all of these groups involves the type of socialization process that is our focus. Perceptions of group variability have also been shown to influence group-relevant judgments (Bogart, 1998; Lambert, 1995; Lambert, Barton, Lickel, & Wells, 1998; Park & Hastie, 1987; Ryan, Judd, & Park, 1996). For example, people who perceive greater variability in the group as a whole are more likely to gather information about a newly encountered individual group member in a way that allows the stereotype to be disconfirmed (Ryan, Bogart, & Vender, 2000). It thus seems important to determine how perceptions of group variability develop. Finally, perceptions of group variability and positivity clearly change during new group membership. Given the importance of stereotypes in our daily lives, it seems imperative that researchers consider in more detail the role that socialization processes may play in stereotype development. Such research is likely to provide a more complete understanding of group socialization, stereotype change and development, and the influence of group socialization on perceptions of groups. We believe that such work will also help identify ways in which groups can promote more positive and constructive group and intergroup perceptions. NOTES 1. Aronson and Mills (1959) are frequently cited as demonstrating that new group members who undergo more severe initiations come to evaluate their group more positively. However, the type of group they examined seems unlikely to have been considered important to participants. Indeed, to maximize cognitive dissonance, the group discussion that participants heard was designed to be “one of the most worth less and uninteresting discussions imaginable” (p. 179). Moreover, participants had no contact with group members, and were thus never treated negatively by group members nor were they led to believe that their negative experiences had anything to do with group members. Instead, the experimenter told participants that he (i.e., the experimenter) had decided to screen potential group members based on his evaluation of an “embarrassment test” in which participants in the severe condition read obscene words aloud. The experimenter further explained that it was easier for him to screen out members early rather than have to ask them to leave after they joined the group. Finally, as we have indicated, a great deal of other empirical work indicates that harsh initiations result in more negative attitudes toward the group. 2. Hogg (2000a) has also examined the role of uncertainty in group processes. His approach is different from ours in its focus on subjective uncertainty reduction as a motivation underlying the desire to affiliate with social groups. We start from the assumption that new members, by virtue of entering the group, are motivated to gain acceptance; their uncertainty, which they also seek to reduce, concerns whether and how acceptance will be achieved.
AUTHOR NOTE This work was supported by Grant SBR-9620113 from the National Science Foundation and Grant R01 MH62061 from the National Institute of Mental Health. We are grateful to Nathan Hart for creating the webpages; Geoffrey Munro, George Schneider, and Stephanie Muraca for help conducting the experiment; and Markus Brauer, Mick Rothbart, Vincent Yzerbyt, and Jamie McMinn for helpful comments on a previous version of this paper. Correspondence concerning this article should be addressed to Carey S.Ryan, Department of Psychology, University of Nebraska at Omaha, Omaha, Nebraska, 68182–0274. Email may be sent to
[email protected].
16 Entitativity, Group Distinctiveness, and Social Identity Getting and Using Social Structure RUSSELL SPEARS Cardiff University DAAN SCHEEPERS Leiden University JOLANDA JETTEN University of Exeter BERTJAN DOOSJE University of Amsterdam NAOMI ELLEMERS Leiden University TOM POSTMES University of Exeter INTRODUCTION In this chapter we consider how perceptions of entitativity and group distinctiveness provide a link between social perceivers and the social world of intergroup relations. Social perceivers are interested agents who are part of the social field they survey, and not just observers of it. Social identity forms a social psychological bridge that relates us to this social structure, mediating and moderating these group perceptions, but also arising from them. We argue that perceptions of group distinctiveness and entitativity play an important role in creating social identity and putting it to use. There is a danger that we view group perceptions (distinctiveness, entitativity, group homogeneity) primarily as outcomes of social perception, as an end in themselves, restricted to the perceptual realm. This may be interesting in its own terms but the function of these perceptions becomes more apparent when we see their close relation to broader questions of social identity, group goals, and group behavior. One purpose of the present chapter is therefore to integrate these perceptions into the broader flow of group life and to examine the social functions they can serve. We use and extend social identity principles to help explain why entitativity and group distinctiveness arise and how and why they are maintained. Clarifying Constructs, Theory, and Measures What is entitativity? It is a simple but slippery construct. Originally coined by D.T.Campbell (1958), it can be defined most minimally as the extent to which the group in question is perceived to form a “real” entity, or the perceived degree of “groupness.” A number of factors can determine perceptions of entitativity. It can be a product of the internal structure within the group, such as the degree of interdependence between group
220
THE PSYCHOLOGY OF GROUP PERCEPTION
members (cohesion, connectedness, etc.). On this interpretation, there is a clear Lewinian legacy to entitativity. In this volume Rothbart and Park define three further related (but distinct) components of entitativity, rising in strength from weak to strong: (1) common attributes, (2) common goals or intentions, and (3) underlying essence. Common attributes are simply properties (e.g., traits, attitudes) that group members have in common. As we make more explicit below, this refers to a basis of group definition that is different from Lewinian notions of interdependence concerning the relation between group members (similarity does not necessarily imply group structure, nor structure necessarily similarity). Common goals or intentions refer to a more psychological basis for entitativity (related to but also distinct from interdependence between individual group members). Essence can also have a psychological basis but goes further in suggesting some deeper physical or material basis to the group, a grounding in “natural kinds” (see Haslam, this volume). According to all of these principles, entitativity is clearly reality based, moving from surface similarity through psychological to physical properties that justify the attribution of entitativity. It is important at the outset to make clear how entitativity relates to other apparently similar constructs used here, and especially to terms relevant to social identity theory that we draw on in this chapter. Two constructs of central importance here are group homogeneity and group distinctiveness. The easier construct is homogeneity, which is generally defined as a perception of similarity or lack of variability/dispersion within the group on a given dimension. As such, it relates most closely to the weakest determinant of entitativity in Rothbart and Park’s scheme. Although homogeneity can be perceived with respect to psychological properties such as traits, it makes no a priori assumptions about deeper psychological or physical causes of these (surface) similarities. In short, homogeneity can be used as one (albeit imperfect) indicator of entitativity. Judgments of entitativity can, however, cut deeper in terms of implied inferences and attributions. Entitativity is also closely related to the concept of group distinctiveness central to social identity theory (Tajfel & Turner, 1979). It is instructive to spend a little time teasing out the overlap and where the differences lie. First, entitativity and group distinctiveness (and their associated approaches) have much in common. Both concepts refer both to the physical reality and to the perception of the group (sometimes a cause for confusion; see Spears, Jetten, & Scheepers, 2002). Clearly there is a psychological/perceptual dimension to both entitativity and group distinctiveness; otherwise this volume would not be of importance to social psychologists. Entitativity and group distinctiveness are therefore not just objective properties of groups, but also social perceptions or even attributions. One important factor that moves group distinctiveness beyond entitativity per se is that the (in)group is defined as part of the self (a “social identity”). Self-categorization theory similarly emphasizes the “group within the individual” (Turner, 1985). Although researchers of entitativity have more recently addressed the relation of self to include others (Aron, Aron, & Smollan, 1992) and even ingroups (L. Gaertner & Schopler, 1998; Schubert & Otten, in press; Tropp & Wright, 2001), it is not clear that the self is theoretically critical to the conceptualization of entitativity in the way that it is to group distinctiveness within the framework of social identity theory. A further key difference that relates to the question of self-relevance is that the social identity concept of group distinctiveness introduces a clear motivational component not always evident in entitativity approaches (i.e., that do not explicitly implicate the group self; but see also Yzerbyt, Castano, Leyens, & Paladino, 2000). As we shall see below, this motivation manifests itself in different ways in different situations. However, the basic theme is that groups, or at least those people who define themselves in terms of the group, strive to attain and maintain group distinctiveness. In short, from the perspective of social identity, distinctiveness is not just a social perception or social reality, it is also a preferred state of being (hence, this is sometimes referred to as “psychological group distinctiveness”). Group distinctiveness
ENTITATIVITY, GROUP DISTINCTIVENESS, AND SOCIAL IDENTITY
221
therefore implies a dynamic state of equilibrium: lack of group distinctiveness is aversive and should motivate attempts at intergroup differentiation to create or restore it, especially among high identifiers (Spears et al., 2002; Tajfel, 1982). Where entitativity theorists have introduced motives into the picture, these tend to be perceptual (Gestalt) or individual (as in the need for structure or for closure), rather than relating to social self and group level motivations. The motivation to differentiate ingroup self from the outgroup “other” in social identity theory implies a further key difference between entitativity and group distinctiveness. Whereas the definition and determinants of entitativity make it possible to conceptualize the group as a thing in itself, for social identity theory the group primarily becomes distinctive in intergroup relations and thus in comparison with other relevant groups or categories. In short, entitativity can be defined in principle independently of the intergroup context, whereas group distinctiveness has little meaning beyond such a context. This difference is partly a product of how the group is conceived. This is not to say that social identity is never defined “immanently” in terms of the interdependent structure or bonds within a group.1 Nevertheless, group identity becomes salient, and its distinctiveness becomes apparent, by virtue of social comparisons with other groups. Self-categorization theory, which in certain respects attempts to move the group from intergroup relations and into the realm of intragroup processes more generally, nevertheless always maintains this intergroup perspective. Classical intragroup phenomena (e.g., group formation, group salience, group polarization, leadership, etc.) are all theorized with respect to the intergroup as well as the intragroup context. Thus, whereas group entitativity may confer “distinctness” in some absolute sense, distinctiveness requires some further relative separation from the surrounding social context in terms of category identity (optimal distinctiveness theory makes a similar point with respect to group size; see Brewer, 1991). In these terms, whereas a degree of entitativity may be necessary for group distinctiveness, it is not sufficient (e.g., common attributes that define entitativity may also be shared with relevant comparison groups, undermining distinctiveness). Conversely, group distinctiveness may be a source of entitativity, but not the only one. These conceptual considerations then raise the empirical issue of how best to measure entitativity and group distinctiveness and their associated processes. This is more complex in the case of distinctiveness for reasons relating to the preceding distinctions, and specifically (a) its relational nature and (b) its motivated character. We address these issues in turn in relation to relevant measures. First, distinctiveness needs to take into account not just the entitativity within the group, but its relation to a relevant comparison group. This point is illustrated if we take homogeneity or group variability as an indicator of the underlying distinctiveness or entitativity. From the perspective of entitativity, homogeneous groups are simply more entitative than heterogeneous groups. However, in the intergroup context, the relation of the ingroup to outgroup on the underlying dimension is also important. Heterogeneous groups may be just as distinctive as homogeneous groups when neither overlap (Jetten, Spears, & Manstead, 1998). For this reason perceived homogeneity is a more partial or contingent indicator of distinctiveness than it is of entitativity. Recent attempts to develop more visual indicators of entitativity by variations on the Aron et al. scale also often fail to capture the intergroup nature of distinctiveness (e.g., Gaertner & Schopler, 1998; Tropp & Wright, 2001), although recent versions have attempted to incorporate this component of intergroup difference (Schubert & Otten, in press; unfortunately too late for the research reported here). Second, the motivated character of group distinctiveness means that judgments of distinctiveness can both form a perception of underlying distinctiveness, and a response to its perceived absence (Spears et al., 2002; see also Yzerbyt, Corneille, & Estrada, 2001). It is then not always clear from some measures which of these two (i.e., input or response) is being measured. Once again, perceptions of homogeneity, even in relative terms, often do not always provide sufficient information to determine this. Although ratings of
222
THE PSYCHOLOGY OF GROUP PERCEPTION
intergroup differentiation provide more evidence of the relative distinctiveness of the groups, again these do not always distinguish the perceived degree of distinctiveness (the fit or metacontrast, to use the terminology of self-categorization theory) from attempts to affirm distinctiveness in its absence. Pictorial measures of (ingroup vs. outgroup) entitativity encounter the same problem. Differentiation on dimensions unrelated to the distinctiveness feedback, or behavioral discrimination (e.g. in terms of reward allocations), can provide more reliable evidence of the attempts to assert distinctiveness. For this reason we include in our arsenal measures of intergroup differentiation and discrimination, in addition to measures of entitativity and homogeneity, where these can shed light on distinctiveness processes. With these conceptual distinctions and measurement issues in mind, we now proceed to elaborate the relation of group distinctiveness to social identity, the different functions group distinctiveness can serve, and the influence of distinctiveness and entitativity on group behavior such as differentiation and discrimination. “Functions” of Social Identity and Group Distinctiveness Addressing these conceptual distinctions already gives us some important hints about the relation between group distinctiveness and social identity. Group distinctiveness both provides the material for social identity (without group difference we would have little basis for distinctive identities) and motivates a potential response to social contexts lacking in distinctiveness. This dual character of group distinctiveness corresponds closely to the duality of social identity as both “being” and a means of “becoming” (Spears, Jetten, & Doosje, 2001). These aspects are reflected in two definitions of social identity elaborated by Tajfel. Tajfel defined social identity as that part of the self-concept relating to our membership in groups, together with the emotional and value significance attached to this (Tajfel, 1978b, p.63): “being.” In the Katz-Newcomb lectures, he also defined it as an intervening variable in a process of social change, a tool to challenge an unfavorable status quo: “becoming” (Spears et al., 2001). Consistent with the Tajfel’s two definitions we propose that group distinctiveness as a perceived property of the group addresses respectively at least two basic functions of social identity. In a few words, it helps us to define and to do. We refer to these as the identity and instrumental functions respectively. The identity function relates to defining who we are in group terms (and again, at the risk of redundancy, this is equally if not more important when our group does not appear to have distinctiveness from others). The instrumental function relates to the implications of this group identity for behavior and addresses the question of how group goals are facilitated by a distinctive and entitative group. These functions can be seen as responding to aspects of the social structure that provides a basis for defining social identity and a context for social interaction and social action. Group identity is shaped by the social structure (e.g., available social categorizations), and group behavior deriving from this identity is likely to be stimulated and constrained by the social structure (e.g., as a response to the group’s position in a status hierarchy). The role of social structure is therefore important here in two distinct senses: (1) Social structure provides an important basis for perceptions of group distinctiveness, and (2) “the” social structure also forms the theater in which group behavior is motivated and can make a difference. In these terms the identity and instrumental functions can be thought of respectively as “getting” and “using” social structure for the benefit of the group. As we shall try to show, it would be wrong to try to separate these functions any more than analytically as they are often related. We argue that questions of action are intimately bound up with issues of self-definition, and indeed conscious action seems to presuppose a corresponding identity (Reicher, Spears & Postmes, 1995; Spears & Lea, 1994).
ENTITATIVITY, GROUP DISTINCTIVENESS, AND SOCIAL IDENTITY
223
So far, and consistent with our social identity perspective, we have put more emphasis on group distinctiveness than entitativity, closely related though these are. Although group distinctiveness is for our purposes more relevant to defining group identity (or motivating a search for it) than entitativity, entitativity is as important if not more so in addressing the instrumental function relating to group action. To the extent that entitativity, more explicitly than distinctiveness, addresses issues of internal structure and organization within the group, this concept is highly relevant to furthering the goals of the group. For this reason we do not see redundancy between these concepts, and we exploit the “distinctive” properties of each, particularly in addressing instrumental concerns. First, however, we begin by considering the identity function and its consequences for group perceptions and behavior. The identity function: Distinctiveness as means to social self-definition Common sense tells us that groups that are objectively entitative and distinctive are more likely to encourage perceptions of entitativity and distinctiveness, and to evoke intergroup differentiation on the underlying dimension reflecting this fact (see also Oakes, 1987). While this often is the case, we propose that judgments of entitativity can also arise from a lack of social structure and clear iden tity, and indeed as a response to it. To perceive or even “project” group distinctiveness, at least where this seems warranted, provides meaning to the social context. Moreover, where we are implicated as group members, it helps to define our social place therein. Although not always explicitly outlined in social identity theory, these ideas are consistent with some of the early writing of Tajfel (1969; see also Tajfel, 1981b; McGarty, this volume), and concur with the social identity principle that people seek positive distinctiveness for the group they are associated with (Tajfel, 1978b; Tajfel & Turner, 1979; see also Brewer, 1991). Usually this distinctiveness principle is taken to apply to group identities that are already established. We label attempts to (re)affirm an established group identity in the face of distinctiveness threat from a similar comparison group as “reactive distinctiveness” (Spears et al., 2002). However, psychological group distinctiveness also provides an important motivation in the process of group formation where this does not yet exist, a process we refer to as “creative distinctiveness” (Spears et al., 2002). This focus on creative distinctiveness should not be taken to deny the more obvious and straightforward prediction that perceptions of entitativity and distinctiveness flow from “objective” evidence of similarities within or differences between groups, which we term “reflective distinctiveness” (Spears et al., 2002).2 In intergroup terms, reflective distinctiveness follows from the principles of metacontrast and comparative fit, and also accentuation theory (Oakes, 1987; Turner, 1985; see also Tajfel & Wilkes, 1963; and Corneille & Judd, this volume; McGarty; Reynolds et al., this volume). Intergroup differentiation based on intergroup differences reflects a basic reality principle and need not invoke identity or instrumental concerns (or indeed the self). A more instrumentally motivated form of reflective distinctiveness may also occur, however. Entitativity arising from internal group structure may result in a group-favoring response for more goal-directed reasons (serving individual or group interests). For example, some researchers have argued that reciprocity and interdependence within groups can produce entitativity and stimulate group-level effects such as ingroupfavoring strategies (e.g., L.Gaertner & Schopler, 1998). The notion that the subjective feeling of groupness is important to achieving group goals goes back as far as Festinger’s (1950) ideas on group locomotion (see Yzerbyt, Castano, Leyens, & Paladino, 2000). Others have focused on essentialist theories that buttress entitativity within the group and provided a basis for perceived group differences (C.Hoffman & Hurst, 1990; Yzerbyt, Rocher, & Schadron, 1997; Yzerbyt et al., this volume). All these factors may underlie a positive relation between entitativity and ingroup bias. From a social identity perspective our argument, however, is that the absence of a meaningful group identity can also motivate a quest for a distinctive group identity (at least among those committed to the
224
THE PSYCHOLOGY OF GROUP PERCEPTION
idea of the group), addressing the identity function. The question then arises about the factors that predict when creative and reflective distinctiveness principles might outweigh each other. We consider some critical moderator variables that speak to this issue toward the end of this section. First, it is important to present some evidence for creative distinctiveness and the identity function that we claim it serves. In a first study (Spears, Jetten, Arend, Van Norren, Postmes, in prep.) we conducted a fairly simple experiment using the minimal group paradigm. In this study we categorized participants according to their preference for one of two painters on the basis of their ratings of a series of paintings (actually by Klee and Kandinsky, as in the classic studies by Tajfel, Billig, Bundy, & Flament, 1971), although assignment was actually at random. For half the participants we then provided further details about this minimal social categorization designed to make the group more meaningful or distinctive. Specifically, we told participants in the meaningful condition that research had shown that people who preferred painter A tended to be more extroverted, and people who preferred painter B tended to be more introverted (or vice versa: the meaningful identity conditions were counterbalanced). Our prediction was that this extra information would provide the groups a more meaningful basis for group identity, and indeed manipulation checks confirmed that people given a personality saw their group as more meaningful. Importantly, we expected that presence of this group identity should to some extent abrogate the need to differentiate the groups in order to create group distinctiveness on these measures. For the participants in the classical minimal group condition we predicted that the absence of this meaningful identity would result in their trying to create distinctiveness on the available measures (reward allocations and evaluative ratings). This is what we found. Participants in the minimal group condition showed reliable levels of differentiation on reward allocations (total points awarded to the ingroup minus the total points award to the outgroup: see, e.g., Diehl, 1990) and on evaluative ratings. Those in the meaningful conditions did not exhibit reliable differentiation. This finding raises fundamental implications about the minimal group effect. If ingroup bias is eliminated so easily, is there any basis for generalizing discrimination to natural groups that presumably already have meaning? There are a number of responses to this. First, a meaningful identity does not necessarily guarantee the elimination of discrimination, because threatening the distinctiveness of an established identity can also lead to discrimination (“reactive distinctiveness”; Jetten, Spears, & Manstead, 2001; Spears et al., 2002). As an experimental paradigm we would argue that the function of the minimal group studies is not so much to form a basis for generalization to groups in general (scaling up to natural group situations introduces many other factors that can exacerbate conflicts and discrimination), but to isolate causal processes and establish boundary conditions. However, if the provision of meaningful identity (or its acknowledgement) can circumvent discrimination, this is an important lesson that has implications for the multicultural society as well as minimal groups (see, e.g., Hornsey & Hogg, 2000). This initial study gave us some confidence in our line of reasoning. It also provides a new twist to explanations of the minimal group bias. Social identity-based explanations of this effect have tended to focus on the “self-esteem hypothesis,” namely, the idea that differentiation can have the effect of endowing group members with a positive social identity and raising self-esteem (Abrams & Hogg, 1988; Oakes & Turner, 1980). Although the enhancement principle is an important component of social identity theory, it would be a mistake to reduce this theory to a simple enhancement motive. As indicated above, group distinctiveness also plays an important role here (see Spears et al., 2002). These data suggest that differentiation could serve an important identity function by providing a meaningful and distinct group identity that is generally absent in the minimal group context. However, it would be premature to make strong claims about this effect without additional investigation. Further research was conducted in order to rule out alternative explanations (Spears et al., 2002, Study 2). In
ENTITATIVITY, GROUP DISTINCTIVENESS, AND SOCIAL IDENTITY
225
the previous study we measured group certainty (see Hogg, this volume) and this did not vary reliably across conditions or mediate the effect. We measured this again in the present study. We also addressed another explanation for our effect. One possibility arising from the previous design is that ingroup bias was reduced in the meaningful conditions because the information about extraversion or introversion encouraged perceivers to individuate group members (i.e., relative to the minimal conditions). It is well known that individuating the outgroup can undermine ingroup bias (Wilder, 1978), and this is consistent with social identity and self-categorization principles. Individuation shifts perception to the interpersonal pole of the interpersonal-intergroup continuum (Tajfel & Turner, 1979) and from group to personal identity (Turner, 1985). To address this possibility, we utilized a more complex design in which we uncoupled the meaningfulness of the ingroup and the outgroup, and manipulated these independently. In other words, the ingroup identity was meaningful (extroverted or introverted) or not, with the same for the outgroup (the only proviso being that when both groups were meaningful, they were characterized by opposite personality forms). As well as measuring intergroup differentiation (using the Tajfel reward matrices and evaluative trait ratings) as in the earlier study, we also included multi-item measures of individuation (four items, e.g., “I see the members of group A[B] as separate individuals”), entitativity (two items: “I have the feeling that members of group A[B] form a well-defined group”), and certainty (two relating to group certainty and two relating to self-certainty, e.g., “To what extent do you feel certain about your group?”; “To what extent do you feel certain of yourself during the experiment?”). The differentiation results in this study were weaker than in the first study, and more complicated. This may be partly due to the partial success of the manipulation: checks indicated that the manipulation of outgroup meaning (but not ingroup meaning) enhanced perceived ingroup meaning. We interpreted this as a comparative contrast effect: outgroup meaning allowed people to define the identity of the ingroup (i.e., in terms of “what it is not,” so to speak). A somewhat similar pattern was found on the entitativity measure. The meaningful conditions generally produced higher ratings of entitativity, but the perceived entitativity of the outgroup in the condition where the outgroup was meaningful and ingroup minimal was especially high (resulting in a three-way interaction, with ingroup and outgroup entitativity as a repeated measures factor in addition to the two manipulated factors). There was some evidence that people in the most minimal condition did display intergroup differentiation (albeit on only the evaluative ratings and not the reward matrices). However, this effect was stronger for those low in self-perceived prototypicality (which can be taken in this context as an individual measure of the perceived absence of group meaning; see Spears et al., 2002). In short, there is some suggestive evidence here that absence of meaning (albeit derived in comparison to the outgroup) once again stimulates intergroup differentiation that provides group distinctiveness. More important for the interpretation of the earlier study, the results of the individuation measures allowed us to rule out the alternative individuation explanation. Specifically, the meaningful conditions were associated with less and not more individuation of group members. The finding on the entitativity measure also underscored the point that meaningful groups are seen more as a group and not as a collection of individuals. Finally, there was once again no evidence that uncertainty reduction played any role in these effects. To summarize, these studies suggest that a lack of a distinctive group identity can stimulate attempts to create the distinctiveness for the group when this is lacking through available means of differentiation. This research questions any simple assumption that entitativity or group distinctiveness will be positively associated with ingroup bias or intergroup differentiation. This assumption is important to address, because, as we have already noted, other group researchers have made this alternative prediction and found evidence for it. For example, in a recent study L.Gaertner and Schopler (1998) predicted and showed that groups
226
THE PSYCHOLOGY OF GROUP PERCEPTION
made entitative by virtue of interaction and interdependence showed greater ingroup bias than less entitative groups. In some of our own research we have also shown that the greater entitativity could also increase intergroup differentiation. For example, studies by Jetten, Spears, and Manstead (1998) showed that a critical degree of group distinctiveness produced greatest intergroup differentiation and that the relation between entitativity/distinctiveness and differentiation maybe nonlinear (see also Jetten & Spears, in press; Spears et al., 2002). Moreover, additional research using the “meaning paradigm” developed in the studies described above has also shown that greater meaning could also produce more differentiation (Arend, 1999; Spears, Jetten et al., Study 3). In that study, group members categorized on a minimal basis and told that this criterion was associated with extraversion/ introversion actually exhibited more ingroup bias than those for whom the minimal categorization was not elaborated (Arend, 2000). How do we explain these apparent anomalies? One important difference between the Arend study and the earlier ones is that this study was conducted in the class so that the group members in all conditions were copresent and visible to each other. In contrast, the studies described earlier were run via computers such that group members were isolated in separate cubicles and thus anonymous. It could be that these differences in procedure emphasize different forms of group identity, and that these forms underlie a different relationship between distinctiveness/entitativity and group effects such as intergroup differentiation. Some speculation is in order here. The social psychology of groups is characterized by two different traditions reflecting different conceptualizations of the group. The Lewinian “interdependence” tradition (Lewin, 1948/1971), views the group as a structure comprising the individuals and the interpersonal relations that make up the group (see also footnote 1). Work teams are good examples of this kind of group, and evidence suggests that such small groups are seen as highly entitative (Lickel et al., 2000). By contrast, the social identity and self-categorization tradition focuses more on groups as psychological entities (as well as social structures) and emphasizes a shared identity as the basis for group membership rather than (or, strictly speaking, in addition to) interdependence or interaction. This distinction has been formalized recently in terms of the distinction between respectively “common bond” and “common identity” groups (Prentice, Miller, & Lightdale, 1994; see also Moreland, 1987; Turner, 1982; Wilder & Simon, 1998, for similar distinctions). Although we think this distinction is useful, it is important not to overstate it. We argue that for common bonds to make a psychological group, they still have to be defined in terms of a common identity (and in fact the evidence of Prentice et al., 1994, is consistent with the view that there is no clear dichotomy; see footnote 1). We therefore prefer to view this distinction as forming a continuum in which common bonds can used to be define a psychological group identity. With this distinction in mind it is possible to conceive the groups constituted in the studies by Arend, and by L.Gaertner and Schopler, as forming more interdependence-based groups. L.Gaertner and Schopler are quite explicit about this, and the social interaction in the high entitativity conditions of their research was explicitly designed to foster this interdependence. The copresence of individuals, as well as rendering group definition in terms of personal bonds possible, allows interpersonal interaction, strengthening these bonds (L.Gaertner & Schopler, 1998). Visibility between group members is an important condition that we would argue can emphasize interpersonal bonds between group members and may have facilitated this instantiation of group identity in the Arend study. By contrast, we argue that anonymity can paradoxically help to emphasize a shared identity. It is relevant to note that some social identity theorists have argued that the power of minimal groups derives from the depersonalizing effects of anonymity (Spears, 1995; Tajfel, 1978b). Tajfel has even suggested that minimal groups are “maximal” groups for this reason (see also Diehl, 1990). This might help to explain why
ENTITATIVITY, GROUP DISTINCTIVENESS, AND SOCIAL IDENTITY
227
differentiation in the minimal conditions is reduced when people were copresent and thus visible as in the Arend study. Indeed, this proposed effect of identifiability is a cornerstone of the social identity model of deindividuation effects (SIDE: Reicher et al., 1995; Spears & Lea, 1992; 1994). According to SIDE, anonymity strengthens group salience and group effects. Visibility may therefore reduce group salience and stimulate group definition in terms of the individuals present and the bonds between them. Why then should a group defined more by common bonds lead to greater ingroup bias when the groups are made meaningful? One possibility is that the emphasis on the interpersonal dimension in this kind of group tips the balance away from distinctiveness concerns and toward interdependence concerns. Interdependence explanations for ingroup bias focus on expectations of reciprocity within the ingroup (Gaertner & Insko, 2000; Rabbie, Schot, & Visser, 1989). This does not mean that issues of group definition are unimportant, however. On the contrary, an ingroup definition is necessary to define the limits of reciprocity within the ingroup (L.Gaertner & Insko, 2000; K.Stroebe, Lodewijkx, & Spears, 2003). The presence of a common identity defined by personality type should therefore reinforce the sense of group identity, that defines the bounds of mutual reciprocation. Another way of stating this is to say that in groups defined more in terms of common bonds and interdependence there is a shift from identity concerns (creative distinctiveness) to more instrumental concerns (mutual group/self-interest). This line of reasoning is consistent with the emerging view that there are two separate motives or routes to responses in the minimal group paradigm: ingroup favoritism based on reciprocation based on interdependence and self-interest, and differentiation based on enhancement and/or distinctiveness principles (K.Stroebe et al., 2003; Yamagishi, 2001). The specific pattern of reward allocations provides some suggestive support for this speculation. L.Gaertner and Schopler (1998) found evidence that ingroup favoritism (“max own” on the MAMs, comparable to the maximum ingroup profit strategy on the Tajfel matrices), increased with entitativity. However, the strategy of maximizing the difference in the rewards to ingroup and outgroup (“max rel own” on the MAMs, equivalent to the maximum differentiation strategy on the Tajfel matrices) did not vary reliably with entitativity. We would predict that a differentiation strategy would better serve the creative distinctiveness needs of establishing an identity than simply maximizing ingroup gain (which need not necessarily differentiate the groups). This is because, as described earlier, distinctiveness is more relational and comparative than entitativity. Although this interpretation of our third study is somewhat speculative and awaits a more direct test, some further research of our own (albeit more concerned with social influence than intergroup differentiation) does provide some suggestive support for this proposed mechanism. In this study (Spears & Postmes, 1998) we actually manipulated “group type” directly by means of feedback about the nature of the group. In a computer communication study participants engaged with three other group members in various tasks designed to emphasize either personal bonds with fellow members or their common basis for categorization in terms of the similarity of their views (common identity). Checks indicated that this manipulation was successful. Crossed with the group-type manipulation, we varied whether group members were anonymous (as in the first two studies of Spears, Jetten et al., above), or identifiable by means of an on-screen photo. The latter condition therefore creates the visibility of the Gaertner and Schopler and Arend studies. We predicted that for the common identity group, anonymity would enhance the salience and depersonalization of the group and strengthen group effects (social influence, perception in group terms, etc.) in keeping with the SIDE model (e.g., Lea, Spears & De Groot, 2001; Spears, Lea & Lee, 1990; Postmes, Spears, Sakhel, & De Groot, 2001). However, we predicted a quite different effect for the anonymity manipulation for the group predefined in terms of common bonds. Here we expected that visibility would strengthen the interpersonal bonds that formed the basis for this kind of group, and group
228
THE PSYCHOLOGY OF GROUP PERCEPTION
related effects. Predictions were confirmed, resulting in a crossover interaction on a range of group-related measures. Social influence and measures of group identity (perceived impression in group terms, perceived collectivism) were higher under anonymity for the common identity group, but higher under identifiability for the common bond group (Postmes & Spears, 1999). To summarize, lack of entitativity/distinctiveness can sometimes provide groups with the motive to differentiate themselves and thus gain distinctiveness (creative distinctiveness). However, there is clearly another side to the story, and in other contexts the presence of group meaning or entitativity can form the basis of differentiation and related group-based responses (reflective distinctiveness), often for instrumental reasons. The type of group (common bond vs. common identity) and contextual moderators (e.g., visibility vs. anonymity) may unite to produce these apparently contradictory responses. A further important feature of the interdependence-based groups used in the research by Gaertner and Schopler is that these were task groups geared to achieving certain goals, rather than just being bound by a common identity. This reaffirms the importance of the instrumental function, which we argue can bolster entitative and group-related effects. We now consider this possibility more directly. Identity Versus Instrumental Functions: Entitativity as Being Versus Doing In the preceding section we focused on the identity function of group distinctiveness. However, exceptions to this already raised the importance of instrumental concerns in the relation between distinctiveness/ entitativity and ingroup bias (taken here to include both ingroup favoritism and intergroup differentiation). As we saw above, there is some suggestion that the instrumental function is more operative under the presence than the absence of entitativity. Once a distinctive group identity is established, the entitativity it provides is likely to prove functional for reasons other than defining identity. Entitativity is likely to support the achievement of group goals that require coordinated group action such as competing with other groups (Festinger, 1950; Yzerbyt et al., 2000). The conditions that elicit this function are likely to be quite different to those demonstrating the identity function. Here entitativity or group identity should be to some extent established, and there should be some group project or goal involved. In this section we continue where we left off in the last and try to distinguish conditions of the identity and instrumental functions associated with distinctiveness/entitativity and ingroup bias. Whereas apparently contradictory findings led us to speculate about the importance of anonymity as a critical moderator before, we tackle the differences between these functions head-on in the following studies by manipulating the critical conditions for these functions more directly (Scheepers, Spears, Doosje, & Manstead, 2002). In the first study we examined the effect of the presence or absence of a group goal relating to an impending group competition. A second within-participants factor was formed by the measurement of intergroup differentiation at two times during the experiment—in this case measured by evaluative ratings of ingroup versus outgroup products—once before the introduction of the goal manipulation, and again after this manipulation. This measure provides an index of both differentiation (suited to the identity function) and ingroup favoritism (suited to furthering group goals). Our rationale was that participants in both conditions would display intergroup differentiation at time 1 as a way of creating group distinctiveness, thereby addressing the identity function. Once this was fulfilled then differentiation at time 2 should subside (having already provided distinctiveness). However, we predicted that ingroup bias should persist in the group goal condition because it should help to galvanize the group for the intergroup competition (instrumental function). This should produce an interaction with intergroup differentiation diverging between group-goal versus no-group-goal conditions at time 2.
ENTITATIVITY, GROUP DISTINCTIVENESS, AND SOCIAL IDENTITY
229
This is what we found. Minimal groups were categorized according to painter preferences into synthetic and analytic perceivers and then rated group products (abstract paintings) composed by members from other groups derived from these two categories. Respondents showed positive intergroup differentiation in their ratings as predicted. After the goal/competition manipulation, participants composed abstract paintings themselves and subsequently rated ingroup and outgroup compositions once more. Groups who expected to compete with the other group for a prize displayed increased positive differentiation, whereas those for whom there was no competition displayed no reliable bias. In short, the presence of a differentiated group identity seemed to provide the basis for further ingroup bias, but only if this was coupled to a related group goal. There is a weakness in this experiment, however. It is difficult to argue that it was the acquired entitativity of the group that facilitated differentiation in the second phase, because both groups had the opportunity to differentiate and thus form a distinctive identity. In order to address this, we conducted a second study in which we manipulated the opportunity to differentiate at time 1 as an experimental factor, and again crossed this with a subsequent group competition manipulation. In this 2×2 design we predicted a crossover interaction on a measure of ingroup bias at time 2. Specifically, we expected that groups without a prior opportunity to differentiate and no subsequent group goal would show high levels of ingroup bias in line with the identity function of creative distinctiveness. Groups with a prior opportunity to differentiate, followed by the prospect of intergroup competition, should also differentiate (fulfilling the instrumental function): entitativity as instrumental “means” to further the group competition. The remaining conditions were expected to fall below these levels because the motivation to attain group distinctiveness was either fulfilled (as in the prior differentiation opportunity/no goal condition) or because the basis for group action provided by a distinctive identity/entitativity was lacking (goal present but no prior differentiation opportunity). This is what we found. Using a similar procedure to the previous experiment we manipulated the possibility of discrimination and the prospect of competition. We obtained the predicted interaction of the group product ratings at time 2. Moreover, this crossover interaction was replicated on other related measures: group cohesion, group identification, and collective self-esteem. These measures tap into a series of closely related constructs that we argue are closely associated with the fulfillment of identity and instrumental functions. Identification, cohesion and collective self-esteem are all markers of identity. At the same time, scoring high on these should be predicted to further group solidarity and cooperation necessary to achieved group goals. The measure of group cohesiveness is closest to the concepts of entitativity and distinctiveness, providing evidence that entitativity was furnished where it was absent in the most minimal and meaningless cell of our design (giving a group identity), and bolstered where it was needed for group action in the most instrumental cell. In line with this interpretation, willingness to work for the group was only enhanced in this most instrumental condition, consistent with this instrumental function. Effort contributes less to the distinctive identity of the group than the other variables (see Scheepers et al., 2002). These two studies provide some further evidence for the identity and instrumental functions associated with intergroup differentiation and for the conditions that distinguish between them. We suggest that distinctiveness and entitativity function in two quite different ways in these studies. First, (cre ative) distinctiveness can be an effect or outcome that addresses the lack of distinctive group identity (distinctiveness as “ends”). These two studies provide further evidence that lack of distinctiveness can produce differentiation to supplement the studies in the previous section. Second, once this distinctive identity is provided, the resulting entitativity (solidarity and interdependence within the group) can form a more solid base from which to launch group action (entitativity as “means”).
230
THE PSYCHOLOGY OF GROUP PERCEPTION
These two processes are not separate but complementary. Indeed, distinctiveness as ends is a prerequisite for entitativity/distinctiveness as means. Thus the preceding research shows that a distinctive identity actually enhances the pursuit of instrumental group goals. This helps to make sense of the apparently contradictory findings in the previous section where we showed that an absence of a distinctive identity could sometimes result in more differentiation whereas on other occasions (i.e., after psychological group formation) its presence could also result in differentiation. The studies therefore present evidence illuminating the conditions under which each function can hold, helping to resolve apparent contradictions in the relation between differentiation and distinctiveness/ entitativity. Identity and Instrumental Functions: Being and Doing In the previous section we started to introduce elements of the social context, such as intergroup competition, that influenced the role of differentiation and entitativity as a goal directed function toward certain ends. We contrasted this with the identity function, where identity is not yet formed. Here we move further along the road of group life and consider groups where social identity has to some extent at least been formed. In other words, there is less need for creative distinctiveness. We move on to consider further aspects of the social context (or the social structure) that can affect entitativity. At first sight this shifts our focus more toward the instrumental function of entitativity and distinctiveness, as means rather than ends. Does this mean that the identity function becomes less important? We think not. First, the fact the identity is present does not mean that it is unaffected by what is going on. Identity needs to be protected and maintained. For example, distinctiveness once formed can be threatened (Spears et al., 1997; 2002) and needs to be reasserted (“reactive distinctiveness”: Spears et al., 2002). Second, the question of self-definition is central to providing a springboard for action (Spears & Lea, 1994). Identity determines how we act and react, and so it can be seen as the basis for the instrumental function rather than in opposition to it (as the studies of Scheepers et al., 2002, confirmed). Third, and most critical to this section, even when identity is formed, this does not mean that everyone will define themselves in terms of the group to the same degree. The relevance of the identity function in this section shifts to considering differences in the degree to which group members have internalized the group identity as important. Degree of identification with the group will likely affect how people respond to social concerns that implicate the instrumental function. In other words, we would argue that even when identity is formed to some degree, the identity function never disappears entirely. For this reason it is not always meaningful to separate these although, as we try to show below, where possible it is important to assess these functions and show that they both operate. A central theme of the following studies is therefore to distinguish between people who vary in their commitment to the group identity and show how this moderates reactions to the social structure for both identity and instrumental reasons. Our argument is that high identifiers are more likely to act in ways that preserve or even accentuate group entitativity and distinctiveness, when confronted with features of the social structure that threaten group identity, for both identity and instrumental reasons. We begin by showing that commitment to the group affects how people react to threats to identity in terms of perceived homogeneity (Doosje, Ellemers & Spears, 1995). As we noted earlier, group homogeneity is a less than perfect indicator of both entitativity and group distinctiveness, for it neglects some features of both constructs (see above; this study was conducted before some aspects of the present framework were developed and before more sophisticated measures of entitativity were available). However, for the present purposes we predict that perceiving greater homogeneity in both ingroup and outgroup provides a proxy for group distinctiveness and entitativity, potentially serving both identity and instrumental functions. We then
ENTITATIVITY, GROUP DISTINCTIVENESS, AND SOCIAL IDENTITY
231
proceed to assess further evidence that the resulting effects address identity and instrumental functions, respectively. In the first study we wanted to show that group variability judgments were likely to be affected by threats to group identity but that people with different degrees of identification would deal with these threats in different ways. In earlier work we had shown that when a group is threatened by a negative comparison with another group because of inferior performance, one “social creativity” coping strategy that can be used is to accentuate the variability within the groups (Doosje, Spears & Koomen, 1995). In this way the disadvantageous differences are minimized and may not even appear reliable in the statistical sense (“they may seem better than us, but the differences within are greater than the differences between”). This strategy can also serve to protect individual identity (“if my group is variable, I personally may still be quite competent”; see Doosje, Spears, & Koomen, 1995). We suspected that this was not the whole story, however. One disadvantage of this strategy from a social identity perspective is that while addressing the problem of negative evaluation and self-protection, it also tends quite literally to blur the differences between the groups, thereby undermining the distinctiveness of group identity. Put in other terms, it threatens the identity function. We reasoned that this might not be a problem for all group members, but might be for those committed to some degree to their group (high identifiers). Low identifiers should have no problems using the heterogeneity strategy of distancing themselves from the group, or undermining the group distinction, because by definition they tend to value this group identity less. In other words we expected group identification to moderate the reaction to identity threat in terms of perceived group variability. We investigated this hypothesis in two studies, and both supported this hypothesis (Doosje, Ellemers & Spears, 1995). For example, in the second study we created minimal groups and manipulated group identification experimentally by means of a bogus-pipeline procedure. Specifically, participants were fitted up with electrodes attached to the nonwriting hand and told that this measured their “galvanic skin response,” which together with their responses to a series of group-relevant questions allowed us to assess their degree of identification with the group. They then received feedback from the computer that they either scored low or high on group identification (checks indicated the manipulation was successful), later followed by feedback that the ingroup had performed better or worse on a group task than the outgroup (the threat manipulation). We then measured perceived group variability for both ingroup and outgroup on a number of dimensions (one related to the status feedback and four alternative dimensions) by means of 100point-range estimates. As predicted, we obtained a consistent two-way interaction. High and low identifiers did not differ in their variability estimates under low identity threat (superior ingroup performance). However, whereas low identifiers accentuated variability under high threat (inferior performance), high identifiers actually accentuated group homogeneity under these conditions. This pattern was identical for both ingroup and outgroup measures (i.e., there was no evidence of an outgroup homogeneity effect here; ingroups and outgroups were perceived as equally homogeneous). In theoretical terms, one could liken the heterogeneity strategy of the low identifiers to either social mobility or social creativity strategies described by social identity theory (Tajfel & Turner, 1979). It is like a mobility strategy because by emphasizing heterogeneity one emphasizes the distance from those negative people “pulling the group down”, and even enters into outgroup territory if the distributions overlap. It can also be regarded as a social creativity strategy because emphasizing the variability undermines the reliability of the group differences. The strategy of maintaining and even accentuating the group homogeneity shown by high identifiers on the other hand is clearly much more akin to a “social competition” strategy. In psychological terms, these people are not ashamed to be associated with their group and they stick by it. On
232
THE PSYCHOLOGY OF GROUP PERCEPTION
the contrary, if anything, they seem to be affirming group identity just when their group most needs them— when it is under threat. It is important to note that the different pattern for high and low identifiers suggests that this pattern is not produced simply or solely by perceptions of group distinctiveness or entitativity (reflective distinctiveness), but that motivational process play a role here. This is supported by the fact that the same pattern was obtained on the four alternative dimensions on which participants received no status feedback and which therefore provide no perceptual grounds for distinctiveness. So, returning to our distinction between identity and instrumental functions, how should we best characterize the impetus behind the diverging responses of high and low identifiers? On the one hand, it can be seen in terms of an identity function. The low identifiers seem to be protecting their individual self and identity whereas high identifiers seem to be affirming their group self. However, there may also be instrumental concerns at play here. High identifiers, by emphasizing group homogeneity, may be closing ranks, showing a united front, and even preparing for collective action (“united we stand, divided we fall”). The reactions of low identifiers indicate that they protect themselves individually and suggest instrumental concerns: indications are that they would probably skedaddle if they could. We now briefly discuss further studies that point to the operation of both these functions in designs implicating identity threat to both high and low identifiers. Evidence for the Identity Function: Being Evidence that the pattern we obtained on the variability measure in the last study addresses identity concerns is forthcoming from a similar set of studies that employed basically the same design but adopted a different dependent measure: prototypicality ratings or measures of self-stereotyping (Spears, Doosje, & Ellemers, 1997). In these studies we once again threatened group identity, and also distinguished between high and low identifiers. In these studies we threatened identity both in terms of (low) status and also in terms of (low) group distinctiveness, both known to produce threat to group identity (Spears, Doosje, & Ellemers, 1999). As predicted, we obtained a pattern very similar to that obtained for the variability estimates in earlier studies. Specifically, whereas low identifiers tended to distance themselves from the group when identity was threatened by playing down their prototypicality in terms of the group, high identifiers enhanced their prototypicality under conditions of threat. This finding provides more direct evidence that there is an identity dimension to the entitativity strategy. Low identifiers distance themselves from the group, in terms of self-ratings, when the group is negative or indistinct, whereas high identifiers are more likely to see themselves in group terms under these conditions. We are not suggesting that these reactions rule out a link to more instrumental concerns, but they more clearly address the question of self in relation to the group. High identifiers are happy to define themselves in group terms under threat whereas low identifiers are not. In another study (see Spears et al., 1999) we found further evidence for the operation of the identity function in a similar design, on this occasion using as the dependent variable the willingness to categorize the context terms of the group. Specifically, we used the “category confusion paradigm” developed by Taylor, Fiske, Etcoff, and Ruderman (1978). Here we presented the group (psychology students) with statements forthcoming from their own group and a relevant outgroup (Economics students) that depicted either their group, or the outgroup, in a more positive light. Besides measuring the degree of group identification beforehand, we also introduced an additional manipulation designed to increase the salience of group identity. Specifically, in the high salience condition respondents were required to allocate valued
ENTITATIVITY, GROUP DISTINCTIVENESS, AND SOCIAL IDENTITY
233
resources between the ingroup and outgroup using Tajfel reward matrices, a procedure that has been shown to heighten the salience of the social categorization involved (Leyens, Yzerbyt, & Schadron, 1994). As predicted, we found that under identity-threatening conditions low identifiers tended to use the social categorization as an organizing principle to a lesser degree (making relatively fewer within-compared to between-category errors), especially when group identity was made salient. In contrast, high identifiers maintained high levels of social category use under these conditions. In short, whereas under salient identity threat low identifiers shrink from social categorization (and shirk self-categorization), high identifiers do not. Again this indicates two distinctly different orientations to group identity when threat puts the identity question: Who are you? Low identifiers avoid the group, whereas high identifiers embrace it. Evidence for the Instrumental Function: Doing Having shown that the combination of context and commitment speak to identity, what is the evidence that these combinations also evoke instrumental concerns? Support is forthcoming in two studies in which we measured group variability and supplemented this with a more behavioral measure of individual mobility (Ellemers, Spears, & Doosje, 1997). The first study was designed to assess whether people were more likely to want to leave their group when it was threatened or stay behind and remain committed to the intergroup struggle, so to speak. To test this idea we focused on low status groups and added a manipulation of the permeability of group boundaries (Tajfel & Turner, 1979; Ellemers, Van Knippenberg, De Vries, & Wilke, 1988), thereby providing an individual escape route when boundaries were permeable. Once again we distinguished between high and low identifiers by using the bogus pipeline procedure described earlier. Our prediction here was that the low identifiers would be more likely to choose the individual mobility route when they could (permeable group boundaries). We should not have worried about providing an escape route. We actually found a main effect of identification: low identifiers expressed greater preference for the individual mobility option than did the high identifiers irrespective of the permeability manipulation. Moreover, this effect was mediated by the subjective sense of group commitment. High identifiers also emphasized group homogeneity more. Although we should once again be wary of treating this as evidence of either group distinctiveness or entitativity, this measure provides one indicator suggesting that these may have been served. The measure of mobility (or group loyalty) provides more direct evidence that identity concerns also translate into behavior. A second study revealed that low identifiers were more likely than high identifiers to choose for mobility under conditions of high group salience (again by asking participants to make reward allocations between the groups on Tajfel matrices prior to the dependent measures: Leyens et al., 1994). In short, the pattern repeated under conditions in which the group motivations that might be present under conditions of group threat were absent. High identifiers also accentuated group homogeneity under high salience conditions whereas low identifiers tended to emphasize heterogeneity. Once again then, we find a correspondence between the homogeneity measure and the more behavioral index of mobility suggesting that there are material instrumental concerns associated with responses to context, as well as issues of identity. This relation is further illustrated in another study in which we examined the changes in degree of identification over time and group homogeneity judgments in responses to potential changes in the status relations between groups (Doosje, Spears, & Ellemers, 2002). Once again we examined how these reactions were moderated by initial levels of group identification. Here, more explicitly than in previous studies, we were concerned with investigating social identity as an intervening variable in the process of social change, emphasizing the instrumental as well as the identity dimensions to this concept. We created minimal groups
234
THE PSYCHOLOGY OF GROUP PERCEPTION
and then, after they had engaged in a group task, we gave them (false) feedback that their group had performed worse than the other group. We measured their (initial) level of group identification. The groups then performed a second group task and were informed that performance had improved slightly but was still worse than the norm and the other group. However, at this stage a “status stability” manipulation was introduced such that some participants were told that the status relation would in all probability stay the same in future rounds (unlikely change), or that the gap would most likely close and reverse at the current rate (likely change), or that there was a chance of change (uncertain/potential change). We then measured group identification a second time and also gathered group homogeneity estimates (similarity ratings). Our prediction was that identification and homogeneity would increase as a function of the brightness of the group future (from unlikely through potential to likely change). However, crucially we expected this effect to be moderated by the initial level of identification. Specifically, high identifiers should be more prepared to stick by their group under conditions of potential but uncertain change, and would therefore show increased identification and greater group homogeneity under these conditions. This is what we found. Whereas identification and homogeneity did reliably increase as positive change became more certain, this effect was moderated by initial group identification, as predicted. Indeed, high identifiers remained strongly committed to their group irrespective of change levels, whereas low identifiers only showed an increase when positive change was certain. Moreover, high identifiers actually showed greatest homogeneity ratings under the potential change condition. By contrast, low identifiers did not differ reliably in terms of these ratings, although they were somewhat higher for certain change (see Doosje et al., 2002). These findings speak directly to the instrumental function. For the low identifiers, their responses to the social context are individually instrumental in so far as they are only prepared to show increased identification when the future is certain to improve. This is akin to the mobility strategy discussed in the previous studies. The fact that similarity ratings remain low for low identifiers, especially in the unlikely and potential change conditions, is consistent with the individual distancing strategy discussed earlier (Doosje, Spears, & Koomen, 1995; Doosje, Ellemers, & Spears, 1995). The high identifiers show a different strategy but one that is equally instrumental. Identification remains high across outcomes, and group homogeneity is actually highest in the potential change condition. These findings are consistent with the group members turning to support their group in adversity as well as when improvement seems likely. Moreover, the greater homogeneity ratings are indicative of entitative perceptions that focus on the intergroup nature of the situation, and the unity within the ingroup, under conditions when most work is needed to change the unfavorable status quo (and where group effort might reap the most benefit). Like the previous study, we think that these perceptions are not just “mere” perceptions. They are also incipient actions— cognitions that can feed behavioral intentions and promote the appropriate individual or collective behavior that serves preferred levels of identity. To summarize, these studies show that measures of entitativity (group homogeneity) as well as related measures (group identification) and behaviors (mobility versus commitment) point to their instrumental functions of working toward outcomes that serve the dominant identities, be this the individual or the group. However, these functions should not be seen in contrast to the identity function of defining the self. Rather, the instrumental function is premised on the existence of an identity and builds on it. In this sense the identity can to some extent be seen as a prerequisite for the actions and outcomes (Reicher, Spears & Postmes, 1995; Spears & Lea., 1994). At the same time these actions and outcomes also provide identity with its raison d’être.
ENTITATIVITY, GROUP DISTINCTIVENESS, AND SOCIAL IDENTITY
235
CONCLUDING REMARKS We have viewed distinctiveness and entitativity both as a source of meaning (identity function) and as a resource for mobilization (instrumental function) in relation to social identity. These two functions although easily distinguished are not always easily separated. Generating perceptions of entitativity and distinctiveness are part of the process by which identity is formed, and these social identities once formed then help to define our reactions to the social structure. The interaction between these aspects of entitativity is likely to be dialectical. Entitativity provides a means to invest the social self with “a” social structure and provides a base from which to resist and contest aspects of “the” social structure. Although it is important to make the analytic distinction between identity and instrumental functions, we argue that they will tend to go together, quite literally in practice. NOTES 1. This distinction between a group as a thing itself (a self-standing structure) versus relational construct, suggests that the concepts of entitativity and group distinctiveness point to different underlying conceptualization of the group. Relevant here is the distinction made by Prentice, Miller, and Lightdale (1994) between two bases for group membership, mapping onto the interdependence and shared social identity bases of the group, which they refer to respectively as “common bond” and “common identity” groups (see also Durkheim, 1892/1994; Moreland, 1987; and Wilder and Simon, 1998, for related distinctions). While there may indeed be two bases for how the group is defined or comes about, in line with a self-categorization theory we propose that they are united psychologically by their representation as a part of the self. In this respect both common identity and common bonds may form the basis of a group identity. This seems to be supported by the data of Prentice et al. (1994), which showed that an empirical distinction between attachment to the group and to its members holds for common identity groups but not for common bond groups. In our view this is because the attachment to group is defined by attachment to its members in common bond groups. People who only define themselves in terms of attachment to group members but not to the group as a whole are, we would argue, not actually part of a psychological group (qua entity) but are better described in terms of nodes in an interpersonal network (e.g., “friendships”). 2. Yzerbyt, Corneille, and Estrada (2001) indicate a path from the perception of entitativity to inferred essence (comparable to identity here), which is comparable to reflective distinctiveness described here. They also describe the opposite path (from essence to the perception of entitativity), which is similar to our notion of creative distinctiveness.
AUTHOR NOTE The authors would like to thank the editors for their encouragement and also Vincent Yzerbyt, Miles Hewstone, and Emanuele Castano for their invaluable feedback on this chapter. We are grateful to the following for collaboration on some of the research reported here: Susanne Arend, Wim Koomen, Mea van Norren, and Tony Manstead.
17 Social Identity as the Basis of Group Entitativity Elaborating the Case for the “Science of Social Groups Per Se” KATHERINE J.REYNOLDS PENELOPE J.OAKES The Australian National University S.ALEXANDER HASLAM University of Exeter JOHN C.TURNER MICHELLE K.RYAN The Australian National University
The natural knowledge processes with which we are biologically endowed somehow make objects like stones and teacups much more “real” than social groups….” —D.T.Campbell, 1958, p. 17 In his seminal 1958 paper Campbell focused on the question of whether social aggregates such as groups can be considered to be real and have the same status as physical entities. He argued that the domain of psychology is considered more “real” than that of sociology because in the former the biological boundary of the person defines the psychological entity. Where groups are involved there is no clear-cut diagnosed entity. In defense of the position that groups are real, Campbell highlighted four general principles that define when discrete elements are perceived as part of a whole (common fate, similarity, proximity, and pregnance). He then applied these to physical objects (such as rats, stones, teacups) and to social groups. Campbell concluded that the social group can be considered an entity, and this perception therefore opens up the possibility of there being a “separate level of analysis from the biological and psychological” (p. 24) and a “science of social groups per se” (p. 24). While many would agree that psychology is interested in individuals and sociology is interested in groups, understanding the interaction between these two treatments of human nature has been of interest to many social psychologists. Researchers such as Asch, Sherif, Lewin, and Tajfel all believed, like Campbell, that groups are real entities but they concerned themselves with the psychological rather than structural (i.e., similarity, proximity, etc.) properties of groups. In order to answer questions such as “Are groups entities?” “How do we know there is such a thing as a group?” “When does an aggregate of individuals become a group?” these researchers adopted a more social
SOCIAL IDENTITY AS THE BASIS OF GROUP ENTITATIVITY
237
psychological, rather than sociological, approach. Their focus was on the psychological indicators of group membership, including the development of shared attitudes, conformity, stereotype formation, and collective behavior. Although, much research has focused on the external dimensions that Campbell applied to the study of groups, this work has been limited in its contribution to a “science of social groups per se.” There are two main reasons for this. First, the focus on these dimensions alone does not necessarily provide valid and reliable indications that groups are at work in the way people think, feel, and act (Tajfel, 1978a; Turner, 1984). For example, it is possible to have members of the same organization who (a) are dependent on one another in terms of the organization’s continued success (common fate), (b) wear the same uniform (similarity), (c) work in the same area (proximity), and (d) seem to operate in a coordinated fashion (pregnance) but who do not, in any psychological meaningful way, function as a group. In fact, much of the recent interest at the interface between social and organizational psychology is concerned with understanding how to create a sense of “groupiness” and associated loyalty, commitment, and work satisfaction in organizations (S.A.Haslam, 2001)—contexts in which the external conditions of entitativity are commonly evident. We believe that Campbell’s (1958) list is not sufficient because “groupiness” in the fullest (and most interesting) sense is an outcome of collective psychology and therefore depends on the shared categorization of oneself and others as group members. Second, where variables such as common fate and similarity (which are the two factors that have attracted most attention) have been investigated, the underlying assumptions often have been at odds with Campbell’s (1958) assessment that psychology should concern itself with the reality of groups. Much of this work assumes that groups are implicated in psychology only by default (e.g., due to cognitive limitations) and as a result are associated with negative percep tions (e.g., stereotypes) and behaviors (e.g., discrimination) that are characterised by distortion and bias. The focus, therefore, has been on groups as an aberrant force that interferes with the accuracy of perceiving others as individuals. For example, outcomedependence (i.e., common fate) is viewed not as a feature of distinctive and valid group psychology but as a mechanism that can combat group-based influences on impression formation by increasing perceivers’ attentional investment in the process (Fiske & Neuberg, 1990; Ruscher & Fiske, 1990; S.E.Taylor, Fiske, Etcoff, & Ruderman, 1978; cf. Reynolds & Oakes, 2000). In our view, the idea that group-related phenomena are the outcome of psychological deficiencies has contributed not only to a pessimistic view of human experience (Turner, 1999a, 1999b) but has limited the scope, relevance, and vibrancy of social psychology. It is assumed that our in-built psychological (and motivational) limitations lead to group perceptions that in turn drive and perpetuate certain forms of social behavior such as stereotypes, prejudice, ingroup favouritism, and intergroup hostility and conflict (Oakes, Haslam, & Turner, 1994; Sherif, 1967). There is limited recognition of the possibility that it is the political and social reality of group life, first and foremost, that is reflected in group or collective psychology and group behavior. The focus in this chapter is on two social psychological theories, social identity theory (SIT, Tajfel & Turner, 1979) and self-categorization theory (SCT, Turner, 1985; Turner, Hogg, Oakes, Reicher, & Wetherell, 1987), that explicitly recognize the veridicality of groups and seek to understand their psychological implications. Within these approaches (collectively referred to as the social identity approach) it is argued that group memberships, the societal structure that reflects the ebb and flow of various intergroup relations, and collective understandings of those relationships, all are related and together produce social life. We begin with an SIT and SCT analysis of the process through which a group becomes “a group.” It is recognized that surface characteristics do not indicate whether a group has psychological significance for its
238
THE PSYCHOLOGY OF GROUP PERCEPTION
members. In order to know when perceptions, motivations, and goals will be group-based and have the nature of a group entity, it is necessary to understand the categorization process that underlies self-definition as a group member. From the social identity perspective, conditions where social identity is psychologically operative also are those where there will be perceptions of group entitativity and coordinated action in line with the beliefs, values, and goals that define the group. We then discuss social identity theory and the role that the reality of social structure plays in helping to define whether self-other categorization will be individuated or group-based. The subjective social structure defines the context in which groups really operate, and it is this reality that self-categorization processes represent. The emphasis is on the role of power and the nature of intergroup relations (i.e., conflictual or otherwise) on shaping group perception. Two experiments are outlined that examine the relationships between power, social structural dimensions, and social identity and link these variables to group-based perception or entitativity. Entitativity has been defined and assessed in a variety of ways (Brewer & Harasty, 1996; Park & Judd, 1990). In the discussion that follows, entitativity refers to perceiving others (and oneself) as members of social groups rather than in terms of individual characteristics. Based on the social identity approach it is when social identity is salient that impressions of one’s own and other groups are more homogeneous and stereotypic on the dimensions that define the intergroup relationship. Furthermore, behavior should be coordinated in line with the properties that define the group. In line with these points and related work, entitativity was examined in the first experiment in terms of judgments of group homogeneity (dispersion) and stereotypic trait ascriptions (e.g., Park & Judd, 1990), and in the second experiment we included intentions to act in a collective way (e.g., Brewer & Harasty, 1996). Although many variables are related to entitativity, in this chapter we focus on the role of social identity (Lickel et al., 2000; McGarty, Haslam, Hutchenson, & Grace, 1995). This means that, in combination with the other chapters of this book, the reader is exposed to a comprehensive and more detailed examination of the theory and research related to entitativity. It is our view that the social identity perspective provides a detailed and integrated analysis of the psychological processes related to perceptions of entitativity and its implications. SELF-CATEGORIZATION THEORY: SPECIFYING THE PSYCHOLOGY OF THE GROUP It is the cognitive redefinition of the self—from unique attributes and individual differences to shared social category memberships and associated stereotypes—that mediates group behavior. —Turner, 1984, p. 528 The question of what makes a group “a group” has been a focus of SIT and SCT. In order to represent both individual and group behavior, Tajfel proposed a continuum of human behavior framed at one end by interpersonal behavior and at the other by intergroup behavior. The shift along the continuum is associated with distinct forms of social behavior. In particular, it affects levels of uniformity in behavior among group members and their treatment of the outgroup. When a situation is defined in interpersonal terms, there should be variability in behavior toward ingroup and outgroup members. Attitudes and behavior become more grouplike or uniform when the interpretation of the social situation nears the intergroup extreme. In other words, ingroup and outgroup entitativity is positively correlated with movement toward the intergroup end of the continuum.
SOCIAL IDENTITY AS THE BASIS OF GROUP ENTITATIVITY
239
Tajfel’s analysis of group behavior formed the basis of social identity theory (SIT; Tajfel & Turner, 1979), and it remains highly influential. But one aspect that it left underconceptualized was the psychological process that underpins movement along the behavioral continuum. In order to address this issue, Turner (1982, 1984) differentiated between self-description or self-categorization in terms of personal identity and social identity. Shifts in identity elaborated a psychological basis for understanding interpersonal and intergroup behavior. Specifically, the identity mechanism transforms the interpersonal-intergroup continuum into a cognitive, social psychological theory of the group. Building on these ideas, Turner developed SCT (Turner, 1985, Turner, 1987a; Turner, Oakes, Haslam, & McGarty, 1994). At the center of this theory is the idea that the psychological depersonalization of the self in terms of social identity enables intergroup behavior to occur. In this way, the self can be defined in terms of others who exist outside the individual perceiver. It is this process that transforms heterogeneity into homogeneity, individuation into stereotyping, and individual behavior into collective action. Because social life is comprised of groups as well as individuals, the transition from perceptions of ourselves in terms of “I” and “me” to “we” and “us” is viewed as ordinary and normal self-experience. Given this shift, though, it is necessary to address the issue of variability in self-categorization and to discover what determines whether group-based perceptions become psychologically operative. According to SCT, the salient level of perception, individual or group, is dependent on the interaction between a perceiver’s readiness to use particular social categorizations, based on his or her background knowledge and experiences, goals, and motivations (accessibility), and the degree to which there is a match with both comparative and normative aspects of reality (fit). Accessibility or perceiver readiness refers to how “ready” a perceiver is to see a particular situation, event, or stimulus in terms of a certain categorization. Fit relates to how well the stimulus can be accounted for by a given categorization. There are two aspects to category fit—comparative fit and normative fit. Comparative fit is operationalized in terms of the metacontrast principle (Turner, 1985). This principle states that stimuli (including the perceiver) will be categorized as psychologically equivalent when the average differences perceived between them are less than the perceived differences between them and other stimuli included in a given comparative context. Normative fit accounts for the fact that social categorizations must make sense in terms of perceivers’ content-related expectations, theories, and beliefs about the categories being represented (Oakes, 1987; Oakes, Turner, & Haslam, 1991). For example, it is difficult to categorize a class of students in terms of gender if the behavior of the male and female students does not fit the perceiver’s expectations about how males and females should behave. Using principles of accessibility and fit it is possible to predict which of many identities will guide perception and behavior in any given context and the extent to which one’s own and other groups will be perceived as entities. Group perception will prevail when social identity becomes salient—for example, in intergroup contexts where there is a meaningful contrasting outgroup (S.A. Haslam, Oakes, Turner, & McGarty, 1996). However, individuated perception will be more likely to occur if only ingroup members are available for comparison (Oakes et al., 1994, p. 190). In line with these arguments, Reynolds and Oakes (2000) demonstrated that the formation of individuated or stereotypic impressions of a specific target person depended on the nature of the participant-target relationship (interpersonal or intergroup, respectively) rather than on factors considered to increase attentional investment in impression formation such as outcome dependence or accuracy goals (Brewer, 1988; Fiske & Neuberg, 1990; Ruscher & Fiske, 1990; Corneille & Yzerbyt, this volume). It is important to emphasize that the self-categorization process is at all times tied to the reality of the situation. The situation a person confronts influences the level of inclusiveness of categorization and its social meaning (in interaction with the perceiver’s background knowledge, goals, and motives). As an
240
THE PSYCHOLOGY OF GROUP PERCEPTION
example of this point, imagine a female academic having a conversation with a male academic. It is likely that her personal identity will be salient and that individual differences between herself and that other person will be accentuated (relative to the defining ingroup, which in this case could be “academics”). However, if the context is extended to include different others—for example, a representative from university management—categorization of herself and the other academic could be at a higher level as “us academics” who are similar to each other and different from “those managerialists.” Of course, to form such categorizations there would need to be normative fit. Perhaps this would be the case if in discussion the administrator emphasized the virtues of contract academic employment. Given the context-dependent nature of categorization if another male academic joined in the conversation, this same person could categorize herself as “female” and as different from those “males” if there was normative fit in terms of gender (e.g., the males behaved in a stereotypically masculine way). Depending on the salient context, impressions of the very same person may be based on individual characteristics or ingroup or outgroup membership (S.A.Haslam & Turner, 1992; Oakes, Haslam, & Reynolds, 1999). In summary, then, SCT provides a detailed account of the psychological processes that underlie the psychology of the group. It explains how people come to think, feel, and act as a psychological group and, importantly, specifies the circumstances when this will occur. Understanding the self-categorization process explains how and when perceivers define themselves and others as individual and group entities and therefore, when a group is “a group.” Furthermore, it accounts for the context-dependent nature of the selfcategorization process and can explain stability as well as variability in judgments of self and others. SOCIAL STRUCTURE AND SELF-CATEGORIZATION PROCESSES Both theories (SIT and SCT) take for granted and are absolutely committed to the notion that social structure, social context, and society, more broadly, are fundamental to the way that social identity processes come into being, are experienced, and shape cognition and behavior —Turner, 2001, p. xi. Self-categorization processes are inextricably linked to social structural factors, to social reality. While the relationship between social identity processes and social structure is often recognized as being central to SIT, it is believed to be of lesser importance to self-categorization theorists (e.g., Hogg & Williams, 2000; Operario & Fiske, 1999; Sidanius & Pratto, 1999; cf., Turner, 1999b). In part, this could help explain why SCT is often (mis)understood as being concerned only with the cognitive aspects of group membership (but see Turner & Oakes, 1997). It is the case that SCT is an interactionist theory that aims to understand the relationship between the social system and psychological processes (Turner & Oakes, 1986, 1997). Both SIT and SCT emphasize that the reality of group relations is central to the analysis of individuals’ cognition and behavior as individuals and group members. In SIT, it is the social mobility-social change continuum that highlights most explicitly the link between the nature of intergroup relations in the social system and people’s psychology. Social mobility refers to the belief that it is possible to improve one’s position in a social system through exertions as an individual. This belief is consistent with the view that social stratification is legitimate, flexible, and permeable and that it is possible to move relatively freely within that system. Social change beliefs reflect the opposite state of affairs—where one holds the view that to improve or change one’s social position, or in fact to resist change if one is satisfied with the status quo, it is necessary to act collectively with others. Tajfel (1978a) viewed positions on the social mobility-social change continuum to be an outcome of the reality of social structural factors and a powerful determinant of whether attitudes and behavior toward
SOCIAL IDENTITY AS THE BASIS OF GROUP ENTITATIVITY
241
outgroup members would be framed in interpersonal or intergroup terms. For example, where there are “laws, rules, sanctions and social norms designed to prevent the members of a lower status group penetrating into a higher status group” (Tajfel, 1978a, p. 51) and these are judged by low-status group members to be illegitimate and unstable, then “social change” beliefs are likely to develop. Under these conditions there is awareness of cognitive alternatives to the status quo, and this means that conflict or disagreement about the social relations between groups is likely to develop. There is an increased likelihood that group memberships will develop that are related to the social conflict and that group members will act in terms of these identities in order to create or prevent social change. If a social system is permeable and it is believed to be stable and fair, social change beliefs will be less common. It is more unlikely that a social situation will be defined in group-based terms, that people will act in terms of their group memberships, or that certain group memberships and coalitions will develop and become psychologically meaningful. A system also can be viewed as illegitimate but stable or legitimate and unstable. Over time, both of these combinations are likely to lead to increased group behavior and attempts at social change (Tajfel, 1978a). An illegitimate system will eventually be challenged, and instability (through awareness of possible alternatives) will lead to perceptions of illegitimacy. In this way there is a reciprocal relationship between social settings and subjective group memberships. Along with other factors, the relevance of a particular social identity to an individual is likely to increase as a function of the number of social situations where the individual acts in terms of the group membership. As a result, group members are likely to interpret a greater number and variety of social situations as being relevant to their group membership. While SCT is concerned with the determinants of category salience in general, its analysis of when categorization occurs at the intergroup level is consistent with SIT’s analysis of group behavior. In SCT terms the perceived status, stability, legitimacy, and permeability of the social relations between groups are features of the comparative context that influence the beliefs, values, and goals of the perceiver and shape the comparative and normative dimensions of intergroup categorizations. Most obviously, variables that are readily associated with SIT—status, legitimacy, permeability—impact directly on the salience of ingroupoutgroup categorizations (Reicher, Hopkins, & Condor, 1997; Spears, Jetten, & Doosje, 2001). In addition, the SCT variables of perceiver readiness and fit describe more precisely the causal processes that link events in the social system to the psychology of the perceiver. Changes in a person’s background theories and knowledge about the social world as well as variations in the immediate context can lead to different self-other categorizations, which in turn can influence the nature of group behavior, the social structure, and future self-other categorizations (P.Brown & Turner, 2002; see also Reicher, 1996, 2001). Consequently, through SCT we are better able to define and understand the impact that social structural factors have on perceivers’ self-other categorizations as individuals and group members (Haslam, Turner, Oakes, Reynolds, & Doosje, 2002; Oakes et al., 1994; Oakes, Reynolds, Haslam, & Turner, 1999; Reynolds & Oakes, 2000). RECENT EMPIRICAL WORK ON SOCIAL STRUCTURAL DIMENSIONS, GROUP PSYCHOLOGY, AND COLLECTIVE ACTION People who are downtrodden, subjugated or oppressed— formulate their common rejection of the oppressor and join in opposition to him. —Sherif, 1967, p. 2
242
THE PSYCHOLOGY OF GROUP PERCEPTION
In this section we move on to discuss two experiments that explore the interaction between social structural conditions and self-categorization processes in more detail. In particular, we focus on the issue of power, which has received more irregular attention in social psychology than in other social science disciplines (Ng, 1980). This situation is changing because power is central to a growing number of areas, such as the social psychology of organizations (S.A.Haslam, 2001; Hogg & Reid, 2001; Reynolds & Platow, 2003), social dominance theory (Sidanius & Pratto, 1999) and social cognition work on stereotyping and prejudice (Fiske, 1993; Operario & Fiske, 1998). Within this body of work, an unresolved issue is the existence of various definitions of power that often confuse status and power. There clearly is a relationship between status and power in which status stems from power and vice versa (Sachdev & Bourhis, 1991). To confuse the issue further, one’s source of power can emerge from a range of factors (e.g., French & Raven, 1959). Understanding power as a social structural dimension is different in scope from focusing on particular relationships in which a person (or group) has explicit and direct control over outcomes. In the empirical work outlined below, power is manipulated as more of a structural variable (Experiment 1) and as a form of explicit control where those with power could include or exclude others from their group (Experiment 2). In our experiments we were particularly interested in the role of power in framing self-other categorizations at the individual or group level. More specifically, the focus was on understanding (a) the structural conditions under which members of both powerful and powerless groups differentially form group-based judgments and (b) the role such judgments play in group members’ willingness to engage in collective action in the interests of social change. Many researchers suggest that group-based judgments (i.e., stereotypes) are a resource the powerful alone can harness in order to serve their own interests (Jost & Banaji, 1994). For example, Jost and Banaji (1994) suggest that those with power can control ideas, beliefs, and stereotypes in the same way as they control other social and material resources and can thereby instill a “false consciousness” in the powerless such that they become complicit in their own disadvantage. Specifically, disadvantaged groups are supposed to “ascribe to themselves and others traits which are consonant with their social position…rather than question the order or legitimacy of the system which produced such an arrangement” (1994, p. 11). In recent social cognition work, Fiske and colleagues have argued that those with power are more likely to stereotype the powerless than vice versa (Fiske, 1993; Operario & Fiske, 1999). Following their continuum model (Fiske & Neuberg, 1990), individuation results from the attentional investment that flows from a motivation for prediction and control over outcomes (i.e., interdependence; see Guinote, this volume). Where there is asymmetry in power relations, it is the powerless who are outcome-dependent on the powerful and, therefore, motivated to allocate attentional resources to the impression-formation process and form more individuated impressions. The powerful, on the other hand, are too (a) insufficiently motivated to individuate and (b) attentionally overloaded by competing demands to invest any attention in their appraisals of the powerless. They tend, therefore, to form highly stereotypical impressions of those over whom they can exert power (Fiske & Depret, 1996; Fiske, Morling, & Stevens, 1996). Because, in Fiske ’s view, being stereotyped “limits [peoples] freedom and constrains their outcomes, even their lives” (1993, p. 621), this asymmetrical use of the stereotyping process can contribute to social domination and exploitation as much as it emerges from it (see Operario & Fiske, 1998). Jost and Banaji (1994) note that this system-justifying use of the stereotyping process is not an inevitable or unavoidable response to inequality. Likewise, Depret and Fiske (1999) recently recognized that the analysis of power and stereotyping based on the continuum model focuses on an individual-level analysis (see Corneille & Yzerbyt, this volume). They argue that more work needs to be done to understand “when
SOCIAL IDENTITY AS THE BASIS OF GROUP ENTITATIVITY
243
the powerless will accept or challenge the current partition of power” (p. 17). Importantly, we need to know what enables powerless groups to become effective entities, coordinate their thoughts and actions as group members, and instigate social change. Collective resistance requires that the powerless be able to express “who we are” and how “we are different from them,” and the content of such categorizations has to reflect and facilitate a collective challenge to the status quo. Furthermore, powerless and powerful groups may judge the social context and group relations differently, depending on their group interests, goals, and motivations (Brewer & Harasty, 1996; S.A.Haslam et al., 2002; Oakes, Haslam et al., 1999; Oakes, Reynolds et al., 1999). Whether groups see themselves as an entity or not will depend, in part, on the nature of the intergroup relationship and whether it is perceived to be secure or insecure. Under certain conditions the powerless and powerful could define their relationship with the other group as being more ingroup (“we are all in this together”) or as outgroup (“us versus them”). Along these lines, in the first experiment described here, we anticipated that the relationship between the powerful and powerless would be defined in more intergroup (“us” versus “them”) terms by (a) the powerful when relations were secure and permeable and (b) the powerless when relations were insecure and impermeable. When relations between groups are secure and permeable, the powerful achieve their positive distinctiveness and value from being different, distinct, and better than the powerless and tend to maximize the differences between their own group and the powerless group (“us” and “them”). On the other hand, under these conditions the powerless attempt to minimize the differences between their own group and those with power in the interests of achieving desired outcomes (i.e., psychological and material resources possessed by the powerful) through individual mobility strategies (“we are all the same”; Tajfel & Turner, 1979). An extension of this analysis is that the reverse state of affairs could characterize relations between the powerful and the powerless. Social identity theory would suggest that when the relationships between groups are insecure and impermeable, it is the powerless who have more to gain from perceiving the situation in intergroup terms because they are in a position to benefit from a change in the status quo. At the same time, for the powerful, threats to the status-quo may initially invite more intragroup characterizations (e.g., “we share the same interests”). In order to maintain the privileges of power, the powerful in a potentially conflictual situation could initially try to diminish the perception of difference between themselves and the powerless and thereby minimize the basis for conflict. As competition escalates, though, they too may adopt explicitly intergroup categorizations that serve system maintenance (Tajfel & Turner, 1979). Experiment 1: When Do the Powerless Stereotype the Powerful and Form More Entitative Judgments of Their Own and the Powerful Group? Much of the work that has examined responses to power has done so in the context of stable intergroup relations. Predictions regarding insecure group relations have not received the same amount of attention (Fiske, 1993; Depret & Fiske, 1999; Ellemers, Doosje, van Knippenberg, & Wilke, 1992; Ellemers, van Knippenberg, de Vries, & Wilke, 1988; see also Turner & Brown, 1978). In this experiment we focused on unstable group relations where participants were allocated to positions of differential power in an organizational context (Reynolds, Oakes, Haslam, Spears, & Wegener, 1998; Reynolds, Oakes, & Haslam, 1999). They were informed that they either had a powerful position (Level A) or a powerless position (Level C) in a hypothetical organization called Delta Micro Systems. A video was then shown that provided background information about the company. An organizational psychologist told them that the boundaries
244
THE PSYCHOLOGY OF GROUP PERCEPTION
between groups were impermeable and there had been conflict between those at different levels in the organization. After watching the video, participants completed a questionnaire that contained measures of social identity salience and entitativity—levels of identification with people at Level A and Level C in the organization, evaluations of homogeneity, and the favorability of impressions of these groups. Initially, though, participants were asked how much power they thought people at Level A had and how much power they thought people at Level C had. Even though there was conflict between the groups, all participants agreed that those at Level A had more power in the organization than those at Level C. In relation to identification, results indicated that those with more power in the organization identified with people at both Level A and Level C. However, those with less power identified more with people at their own level in the organization, Level C. Also, the powerless formed more homogeneous impressions of both groups compared to the powerful, indicating stronger intergroup categorization on the part of the powerless. In addition, as can be seen in Figure 17.1, the powerless formed much less favorable impressions of people at Level A compared to people at Level C. However, the powerful did not differentiate between the different groups in the organization as much in terms of favorability. Interestingly, members of the powerful group tended to form more favorable impressions of people at Level C than at Level A. Such results indicate that forms of appeasement could be at work whereby the powerful show outgroup favoritism in their impressions of the powerless. Taken together, the results on these three measures provide no evidence to suggest that the powerful form more stereotypic impressions of the powerless than the powerless do of the powerful. In addition, in this study we have clear evidence that the structural relationship between groups in terms of power impacts on whether more intergroup or interpersonal categorizations characterize relations between the powerful and powerless. Most important, this work demonstrates that structural factors can impact on groups differentially depending on their location in the social system and their associated goals and motivations. Compared to the powerful, powerless group members identified more with their own group (identification), formed less favorable impressions of the powerful group (stereotypes), and had more uniformity in their impressions about their own and the comparison group (homogeneity). When the relationship between groups was insecure, it was the powerless who formed more group-based impressions of their own and the powerful group. Other analyses of power (e.g., Mulder, 1977) would have difficulty accounting for findings like those in this study, where the powerless tend to turn away from membership of the more powerful group and form negative impressions of those with power. Part of the problem is that many theories related to power assume that the existing power relations are relatively stable (Sidanius & Pratto, 1999) and that powerless people will adopt an individual strategy in striving for a greater share of this power (S.A.Haslam, 2001; Ng, 1980; Tajfel & Turner, 1979). The idea that group processes can be a vehicle to change the power structure is widely overlooked. Experiment 2: The Impact of Permeability, Stereotyping and Entitative Judgments on Willingness to Engage in Collective Action In order to explore this relationship between group relations and stereotyping and collective action (both indicators of entitativity) further, we conducted an additional study that elaborated upon recent work by Wright and colleagues (Wright, Taylor, & Moghaddam, 1990; see also Reicher, 1996; Reynolds, Oakes, Haslam, Nolan, & Dolnik, 2000). Building on social identity theory, Wright et al. (1990) investigated responses to disadvantage and identified the conditions that influence willingness to engage in collective
SOCIAL IDENTITY AS THE BASIS OF GROUP ENTITATIVITY
245
FIGURE 17.1. Favorableness of impressions of people at Level A and Level C as a function of participants’ position in the organization
action (see also Boen & Vanbeselaere, 1998; S.A.Haslam, 2001; Kelly, 1993; Lalonde & Silverman, 1994; D.Taylor & McKirnan, 1984). The findings revealed that members of low-status groups responded to their inequality in different ways depending on the openness of group boundaries and their own individual performance. Most important for the current discussion, impermeable group boundaries were most likely to provoke attempts to instigate collective action—the strategy most disruptive to the existing power structure. In extending this analysis to entitative thinking by the powerless our aim was to show that the use of stereotypes depends on whether group members accept or reject the power structure. In this way we endeavored to extend Wright et al.’s (1990) analysis and demonstrate that group-based social categorizations and stereotypes are responsive to the perceived nature of group relations and are part of the process of social change. Of particular interest was whether participants reproduced the stereotype of the powerful group that was provided by the experimenter, the stereotype “consonant with…social position” (Jost & Banaji, 1994, p. 11). The powerful group was characterized by creativity, rigorousness, complexity, analytical ability, and conscientiousness. In order to examine these predictions we adapted the procedure used by Wright et al. (1990). All participants were assigned to a low-status group and failed to gain promotion to the high-status group, ostensibly due to either inadequate performance (open condition) or because, despite having performed at the required level, the high-status group used their power to reject their attempt at upward mobility by admitting only 10% of those achieving the required scores or by admitting no one. Thus, the intergroup boundary varied in permeability from completely open through a “tokenist” permeability (see Wright et al., 1990) to completely closed. We predicted that both the action choices and the stereotyping engaged in by disadvantaged group members would vary depending on how they responded to their disadvantage—acceptance, individual mobility, collective action (see also S.A.Haslam et al., 2002; Reicher, 1996; Stott & Drury, 2000). More specifically (and following the findings reported by Wright et al., 1990), we expected the open condition to produce acceptance of disadvantage and fairly consensual stereotyping in terms of traits “consonant with… social position” (Jost & Banaji, 1994, p. 11). The 10% quota condition creates a permeable boundary that should encourage attempts at individual mobility, reduce the tendency to define the context in intergroup terms, and therefore reduce stereotyping. The closed condition, on the other hand, was expected to promote
246
THE PSYCHOLOGY OF GROUP PERCEPTION
Note: ** p<.01, * p<.05 FIGURE 17.2. The role of the nature of intergroup boundaries and the favorability of the outgroup stereotype in explaining powerless group members’ willingness to engage in collective protest
collective protest and to produce strongly stereotypical perception of both groups but not in terms of the existing social order. Under these circumstances disadvantaged group members were expected to use the content of stereotypes to express their rejection of their position and to craft definitions of both ingroup and outgroup that provided a social-psychological basis for unified action against the outgroup (cf., Simon, 1998a; Simon, Glassner-Bayerl, & Stratenwerth, 1991, p. 265). Furthermore, we anticipated stereotyping should be a central, and possibly mediating, feature of endorsement of collective action. Findings in the open and closed conditions of this study clearly confirmed our predictions. Where the group boundary was open we observed acceptance and stereotype reproduction. However, where the boundaries were firmly closed, participants exhibited a preference for collective protest that represented a strong challenge to the existing status relationship, and this challenge was backed up by creative negative stereotyping of the outgroup. More specifically, where group boundaries were closed, there was a clear rejection of the view provided by the experimenter that the powerful group was characterized by creativity, rigorousness, complexity, analytical ability, and conscientiousness. The high-status group was described as “mean,” “cold,” and “rude,” and the term “analytical” was used, but it seemed to take on a different, more negative tone in this context as compared to the “analytical, conscientious, and creative” image that was reproduced in other conditions. Furthermore, the relationship between the permeability of group boundaries and willingness to take collective action was mediated by the favourability of the outgroup stereotype (see Figure 17.2). In other words, the nature of the power less group’s stereotype of the powerful group played a significant role in paving the way for social change. Based on these findings it appears that stereotyping can contribute to processes of social change as well as to the maintenance of current power and status relationships. Where intergroup boundaries were open, our participants reported a stereotype of the powerful group that was entirely consistent with, and therefore in a sense justified, its position of power and influence. Under closed conditions, however, we replicated previous findings of a strong preference for collective action against the powerful outgroup, and extended these findings by demonstrating that stereotyping processes also varied with, and contributed to, this move to challenge the status quo. Clearly, the powerless do stereotype the powerful, and this stereotyping can represent both acceptance and rejection of their power. To summarize, both of the experiments reported here address the interrelationship between social structure and variability in categorization of oneself and others as group members. Entitativity or group
SOCIAL IDENTITY AS THE BASIS OF GROUP ENTITATIVITY
247
perception was most evident under those conditions where the social identity perspective would predict social identity salience and group behavior—that is, where an intergroup “us versus them” relationship was most psychologically meaningful. Furthermore, the pattern of findings was consistent across a range of measures of entitativity (stereotypic trait ascriptions, dispersion or homogeneity, and intention to act collectively). The first experiment revealed that under conditions where the boundaries between groups were conflictual and unstable, it was those with less rather than more power who formed the most homogenous impressions of their own and the powerful group and had the least favorable stereotype of the other group. In the second experiment, it was demonstrated that (a) the content of stereotypes changed as a function of the social structural conditions (open, quota, closed) and (b) stereotyping played a role in group members’ willingness to engage in collective action. Based on this analysis both powerful and powerless group members could behave in more or less variable ways depending on the social context (Guinote, this volume). It seems then, that the degree to which one’s own and other groups are perceived as a collective entity is not only influenced by social structural conditions that encourage intergroup categorizations (power asymmetry, insecure group relations, impermeability) but, under certain conditions, can enable the structure itself to be challenged. CONCLUSIONS The fact that human beings are able to act as both individuals and group members is a plus, adding immensely to the sophistication and possibilities of our social relationships. —Turner 2001, p. xi This chapter started by claiming that groups are real and have a distinct psychological significance for their members. SCT was discussed with particular emphasis on the conditions that shape whether impressions will be individuated or group-based. In addition, the interaction between social structural dimensions and psychological processes was discussed. Variables such as power, status, stability, and permeability shape the social context, the extent to which self-other categorizations will be intergroup and more stereotypicbased or interpersonal and more individuated in nature, and the social meaning of such categorizations. In line with this analysis, two experiments were described that explored the interaction between social structural dimensions and self-categorization processes in more detail. The results suggest that certain forms of intergroup relations impact on self-categorization as a group member and therefore group perception and collective action. In Experiment 1, in line with predictions, when group relations were insecure and impermeable, group identification, stereotyping, and judgments of homogeneity increased for members of the powerless, more than for the powerful, groups. Furthermore, results from the second study indicated that under conditions where group boundaries were impermeable rather than open, those with less power formed more negative stereotypes of the powerful outgroup. And the results showed that it was the ability to stereotype in this way that mediated the willingness of the powerless to organize a collective plan of action with a view to bringing about social change. Our results support the view that social structural dimensions do impact on the self-categorization process. These factors shape social reality and, therefore, affect both the extent to which it is psychologically meaningful to categorize oneself and others on an interpersonal or intergroup basis and the content of such categorizations. The conditions that define social conflict from the perspective of low status groups in SIT, impermeability and insecure relations (i.e., perceptions of illegitimacy and instability), are those that increase the salience of intergroup categorizations.
248
THE PSYCHOLOGY OF GROUP PERCEPTION
Furthermore, there is evidence that the subjective social structure affects the social meaning of ingroupoutgroup categorizations. Under certain condition, ingroup and outgroup stereotypes serve to justify rejection of the status quo. Self-categorization processes do not just reflect current features of social life; they can emerge to challenge and change the nature of intergroup relations (Reicher, 1996, 2001; S.A.Haslam et al., 2002). Entitativity is a part of this process as groups generate stereotype content that defines and justifies the challenge. Shared attitudes and behavior are a central feature of social life and social change. In conclusion, it is our view that in order to address the vast array of expressions of group psychology it is necessary to understand fully the psychological implications of a social reality that comprises groups as well as individuals. Too much research time has been devoted to the assumption that our analyses of group psychology only need to account for those (negative) aspects of social life that we may disagree with and seek to minimize. There is a risk that this has blinded theory and research, and social psychology in general, to the vast array of outcomes of group psychology and the potential of such relationships. Our theoretical and empirical work has to be able to explain and be reconciled with those aspects of group life we may seek to minimize (e.g., discrimination, hostility, domination) but it must also deal with those we are keen to encourage (e.g., tolerance, removal of oppression, unified action and cooperation). The SCT analysis attempts to identify the underlying psychological process related to these dynamics and as a result advances in the fullest possible sense the “science of social groups per se” that Campbell encouraged us to aspire to.
18 The Perception of Outgroup Threat Content and Activation of the Outgroup Schema TIM WILDSCHUT University of Southampton CHESTER A.INSKO University of North Carolina at Chapel Hill BRAD PINTER University of Washington
One of the enduring questions in social science relates to whether decent individuals are prone to behave indecently when banded together in a group. Around the turn of the twentieth century, Le Bon (1895) formulated the first systematic analysis of this problem. He proposed that shared membership in a group puts people “in possession of a sort of collective mind” (p. 27). This collective mind was thought to be primitive, irrational, and destructive. Floyd Allport (1924) is well known for his criticism of Le Bon’s crowd mind concept. Nevertheless, in his later writings, Allport proposed that “even if we got rid of the crowd mind, the problem of describing the differential of crowd-like behavior…would remain” (1962, p. 6). In an attempt to determine when and why intergroup relations are more antagonistic and competitive than interindividual relations, research on interindividual-intergroup discontinuity has studied the “differential of crowd-like behavior” in the context of mixed-motive matrix games, like the prisoners dilemma game (PDG). This research was guided by the assumption that the greater competitiveness of groups arises from a confluence of mechanisms that can be studied empirically rather than from the somewhat esoteric crowd mind postulated by Le Bon (1895) (e.g., Insko, Schopler, Hoyle, Dardis, & Graetz, 1990, Insko et al., 1987, 1988, 1992, 1993, 1994, 1998, 2001; McCallum et al., 1985; Schopler, Insko, Graetz, Drigotas, & Smith, 1991; Schopler et al., 1993, 1994, 1995, 2001). The typical experiment in this research program was conducted in a suite in which individuals or groups were located in different home-rooms that were connected to a central room. After examining a version of the PDG matrix provided for a given trial, individuals or group representatives (or in some instances entire groups) went to the central room to discuss possible action with their opponent (or opponents) and then returned to their home-rooms, where they made a choice. In the intergroup condition, the choice was made by the group as a whole. Although there were a number of studies involving just one trial, typically the procedure was repeated for 10 trials. Studies following this procedure have typically found a significant and descriptively large interindividual-intergroup discontinuity effect, that is, a tendency in the context of
250
THE PSYCHOLOGY OF GROUP PERCEPTION
mixed-motive situations for intergroup interactions to be more competitive and less cooperative than interindividual interactions (Wildschut, Pinter, Vevea, Insko, & Schopler, 2003). Given the robustness of the discontinuity effect, we suspect that it is multiply determined. Three hypotheses have been advanced to account for the discontinuity effect. The schema-based distrust (or fear) hypothesis is based on the tendency to distrust other groups more than other individuals. It proposes that intergroup interactions tend to be more competitive than interindividual interactions because the anticipation of interacting with another group activates a negative outgroup schema, consisting of learned beliefs that other groups are competitive, deceitful, untrustworthy, and hostile (Insko et al., 1990; Insko & Schopler, 1998). The social support hypothesis is based on the fact that group members can provide each other with social support for the pursuit of immediate self-interest, in contrast to the lack of such social support available to individuals (Wildschut, Insko, & Gaertner, 2002). Finally, the identifiability hypothesis is based on the person’s assumption that, in an interindividual interaction, he or she can be identified and held accountable for self-interested behavior, and the person’s assumption that in an intergroup setting the opponent’s ability to assign personal responsibility for self-interested behavior is somewhat limited. Stated more succinctly, group membership provides a shield of anonymity (Schopler et al., 1995). In the remainder of this chapter, we review past and present research on the schema-based distrust, or fear, explanation of the interindividual-intergroup discontinuity effect. Note that in postulating the existence of general distrust of outgroups, we are not articulating a new idea. Campbell (1967), for instance, argued that “if most or all groups are in fact ethnocentric, then it becomes an ‘accurate’ stereotype to accuse an outgroup of some aspects of ethnocentrism” (p. 823). Campbell went on to describe a “universal” outgroup stereotype, “of which each ingroup might accuse each outgroup, or some outgroup, or the average outgroup” (p. 823). From the perspective of the observer, this universal stereotype of the outgroup includes dishonesty, competitiveness, and hostility toward the observer’s ingroup. An interesting historical illustration of such schema-like expectations occurs in Plato’s Republic (1891), where Polemarchus defends the principle that “justice is the art which gives good to friends and evil to enemies” (p. 7). Highly reminiscent of this traditional maxim of Greek morality, Tajfel (1970) interpreted the ingroup favoritism demonstrated in early minimal group paradigm experiments as flowing from a learned “‘generic norm’ of behavior toward outgroups,” a norm dictating that we “act in a manner that discriminates against the outgroup and favors the ingroup” (pp. 98–99). This chapter is divided into two parts. The first part is concerned with the content of the outgroup schema. It highlights six categories of evidence indicating that the outgroup schema consists of learned beliefs and expectations that other groups are competitive, deceitful, untrustworthy, and hostile (see also Insko & Schopler, 1998). The second part examines the antecedents of outgroup, schema activation. THE OUTGROUP SCHEMA: CONTENT Discontinuity with the PDG-Alt Matrix The PDG involves an interaction between two sides (individuals or groups), usually over monetary outcomes. Each side can choose between a cooperative (X) and a noncooperative or competitive (Z) choice, and both sides’ outcomes are determined by the combination of their respective choices. A sample PDG matrix is presented in Figure 18.1. That this is a PDG matrix can be verified by noting that the outcomes of the column player, for example, decrease in rank order across cells from the upper right (ZX) to upper left (XX) to lower right (ZZ) to lower left (XZ). Note further that the average outcome in the lower left (XZ) and upper right (ZX) cells is lower than the outcome in the upper left cell (XX). This requirement guarantees that
THE PERCEPTION OF OUTGROUP THREAT
251
FIGURE 18.1. A PDG matrix
the players cannot receive higher outcomes by taking turns selecting X and Z than by both selecting X. The dilemma faced by both sides is reflected by the fact that, on any given trial, each side can maximize its outcomes by selecting the competitive Z alternative regardless of the alternative selected by the other side. Yet, paradoxically, when both sides select Z, both achieve outcomes that are lower than the outcomes they could have achieved by mutual X selections. As Ridley (1996) points out: “Broadly speaking any situation in which you are tempted to do something, but know it would be a great mistake if everybody did the same thing is likely to be a prisoner’s dilemma” (pp. 55–56). Given this outcome array, why might a player select Z, the competitive choice, instead of X, the cooperative choice? There are at least two possible motives. One motive for selecting Z is the self-interest, or greed, that is associated with receiving the highest possible outcome (60 for the matrix in Figure 18.1). Another motive is the fear of receiving the lowest possible outcome (30 for the matrix in Figure 18.1). Theoretically, greed is based on the expectation that the opponent will choose X, and is therefore vulnerable to exploitation. Fear, on the other hand, is based on the expectation that the opponent will choose Z, and therefore poses a threat. In terms of interdependence theory (Kelley & Thibaut, 1978), self-interest, or greed, refers to a concern for maximizing one’s own absolute outcomes (max own), a concern for maximizing one’s own outcomes relative to the outcomes of the opponent (max rel), or both. Similarly, fear refers to a concern for avoiding minimal outcomes in an absolute sense, a concern for avoiding outcomes that are lower than the outcomes of the opponent, or both. Thus, the distinction between max own and max rel is orthogonal to the distinction between greed and fear. This is important because it demonstrates that greed and fear can refer to a concern with the tangible outcomes emphasized by Campbell (1965) in realistic conflict theory, the relativistic outcomes emphasized by Tajfel and Turner (1979) in social identity theory, or both (see Insko et al., 1992, for an empirical investigation of this issue). The fact that the Z choice in a two-choice PDG confounds fear and greed interferes with the goal of studying the independent contributions of fear and greed to interindividual-intergroup discontinuity. To avoid the confounding of fear and greed, Insko et al. (1990) developed a three-choice matrix in which each player has an alternative to either cooperating or competing. An example of this PDG-Alt matrix is presented in Figure 18.2. The four corner-cells of this matrix constitute a PDG matrix. For the remaining cells, which make up the center column and row of the matrix, the outcomes for both players are exactly intermediate to the outcomes in the upper-left cell and the outcomes in the lower-right cell. On the PDG-Alt, choosing Y results in intermediate outcomes regardless of the opponents choice, and is therefore referred to as
252
THE PSYCHOLOGY OF GROUP PERCEPTION
FIGURE 18.2. A PDG-Alt matrix
“withdrawal.” Given the expectation that the opponent is competitive and will choose Z, the Y choice maximizes own outcomes. Given the expectation that the opponent is cooperative and will choose X, however, the Z choice maximizes own outcomes. If a player chooses Z on the PDG-Alt matrix, and thus avoids the safe Y choice, this likely reflects an attempt to maximize self-interest, or greed, at the expense of a vulnerable opponent. The first category of evidence for the outgroup schema comes from four studies on interindividualintergroup differences within the context of the PDG-Alt (Insko et al., 1990, 1993; Schopler et al., 1993, 1995). The schema-based distrust, or fear, hypothesis predicts that on the PDG-Alt matrix groups will make more withdrawal choices than will individuals. The social support and identifiability hypotheses predict that, because groups are greedier than individuals, groups will compete more than individuals even when, as on the PDG-Alt matrix, a withdrawal option is available. Consistent with the idea that the discontinuity effect is rooted in the greater fear and greed in intergroup relative to interindividual interactions, all of the above-referenced PDG-Alt studies found that groups made significantly more withdrawal and competitive choices than did individuals. Given our present concern with the perception of outgroup threat, the study conducted by Insko et al. (1993) is of special interest. This study investi gated the effect of communication on interindividualintergroup discontinuity within the context of a PDG-Alt matrix. Before summarizing the results of the study, we will briefly introduce its theoretical basis. As mentioned earlier, interindividual-intergroup discontinuity has typically been examined within the context of mixed-motive matrix games in which communication between players was allowed. In some studies the communication between groups involved group representatives (e.g., Schopler et al., 1993), in some studies the communication involved a meeting between all members of both groups (e.g., Insko et al., 1987), and in some studies the communication involved the use of an intercom system (e.g., Insko et al., 1993). A frequent outcome of these discussions between players was an agreement to cooperate on the upcoming trial, or at least an assertion by one side or
THE PERCEPTION OF OUTGROUP THREAT
253
the other that they would cooperate on the upcoming trial. For example, coding of tape-recorded intercom discussions in a one-trial experiment with the PDG-Alt matrix indicated that there was a consensus to cooperate in 85% of the intergroup sessions and 76% of the interindividual sessions (Schopler et al., 1995). The schema-based distrust hypothesis postulates a schema consisting of learned beliefs that intergroup relations are competitive, unfriendly, deceitful, and aggressive, which dictates distrust of outgroups. The influential program of research on attitudes and communication conducted at Yale University in the 1950s identified trustworthiness as an important component of source credibility (Hovland, Janis, & Kelley, 1953), and subsequent research has demonstrated that the trustworthiness of a source affects the persuasiveness of communication (Walster, Aronson, & Abrahams, 1966). This implies that, if there is a schema dictating distrust of an outgroup, communication between groups should be less credible and persuasive than communication between individuals. Hence, the presence or absence of communication between groups should have a smaller impact on cooperation than should the presence or absence of communication between individuals. In other words, the schema-based distrust hypothesis implies that communication between individuals should produce a greater increase in cooperation than should communication between groups. Insko et al. (1993) tested the predicted groups versus individuals by communication versus nocommunication interaction effect on cooperative choices within the context of a one-trial experiment with the PDG-Alt matrix. Communication was manipulated by varying whether individuals and groups did or did not have the opportunity to communicate with the opponent through an intercom system. Results indicated that the critical individuals versus groups by communication versus no communication interaction on cooperation was significant. Consistent with the schema-based distrust, or fear, hypothesis, communication produced a greater increase in cooperation between individuals than between groups. In fact, the descriptively small effect of communication in intergroup interactions was not significant. Cooperative intentions communicated between groups did not increase cooperation, presumably because neither group perceived the other group’s message as credible. Explicit Beliefs and Implicit Associations Regarding Intergroup Interactions We define a schema as a set of beliefs serving to organize and guide memory for past events and expectations regarding future events. The above-discussed results are consistent with the idea that within the context of mixed-motive matrix games there is more distrust between groups than between individuals. The second category of evidence in fact demonstrates that people have conscious beliefs and unconscious associations indicating that outgroups are competitive and untrustworthy. A study by Hoyle, Pinkley, and Insko (1989; Study 1) examined the difference between college students’ perceptions of interpersonal and intergroup social interactions. Participants rated the extent to which 50 trait adjectives characterized typical interindividual interactions or, in a different condition, typical intergroup interactions. Participants were given illustrations of interindividual or intergroup interactions (e.g., two students in a particular class, two fraternity groups). Factor analysis yielded two seven-item scales. One scale, labeled agreeableness, was defined by trait adjectives such as cooperative, trustworthy, and helpful. The other scale, labeled abrasiveness, was defined by trait adjectives such as competitive, boastful, and domineering. Results indicated that intergroup interactions were perceived as less agreeable and more abrasive than were interindividual interactions. Hoyle et al. (1989) measured participants’ beliefs about interindividual and intergroup interactions using traditional self-report ratings. More recently, the importance of implicit processes in social cognition has been emphasized. Greenwald and Banaji (1995) define implicit attitudes as “introspectively unidentified (or
254
THE PSYCHOLOGY OF GROUP PERCEPTION
inaccurately identified) traces of past experience that mediate favorable or unfavorable feeling, thought, or action toward social objects” (p. 8). If the outgroup schema is rooted in people’s differentially negative experiences with outgroups, traces of those experiences, in the form of implicit associations, should be observable on measures which do not require participants’ awareness or motivation. This hypothesis was investigated by means of the Implicit Association Test (IAT; Greenwald, McGhee, and Schwartz, 1998), a widely used technique for measuring implicit attitudes. The IAT assesses the relative strength of pairs of associations within the context of a word classification task. Participants classify words into appropriate categories using two response keys (e.g., ‘d’ and ‘k’). For the critical trials, each response key corresponds to two categories (e.g., ‘d’ for FLOWER/PLEASANT and ‘k’ for INSECT/UNPLEASANT). The relative strength of association between category pairs is measured by comparing the average response time for word classifications under one configuration of category pairs (FLOWER/PLEASANT and INSECT/ UNPLEASANT) to the average response time under the other configuration of category pairs (INSECT/ PLEASANT and FLOWER/UNPLEASANT). For example, if people have more positive and less negative implicit attitudes about flowers than about insects, then we would expect the average response time to be shorter under the former configuration than under the latter configuration. This prediction was, in fact, confirmed by Greenwald et al. (1998). In a series of experiments, Pinter and Insko (2003) investigated implicit and explicit attitudes toward individuals and groups. In one experiment, participants were presented with two IATs. One IAT was designed to assess feelings about individuals and groups using a general evaluative dimension (i.e., pleasantness/unpleasantness). Words were selected for the classification task based on typicality of the categories they represented (e.g., one, single, many, collective, diamond, peace, war, evil). The second IAT was designed to assess relative beliefs about individuals and groups, and included agreeableness and abrasiveness adjectives used previously by Hoyle et al. (1989). Results from both IATs were consistent with the existence of a negative outgroup schema. Participants more quickly classified words when the response keys paired words representing the individual and group categories with words representing the pleasant and unpleasant categories, respectively. In addition, explicit measures showed the expected results favoring individuals and were related to the implicit measures. Two additional experiments used an alternative response latency measure, the Go/No-Go Association Task (GNAT; Nosek & Banaji, 2001), to replicate and extend the preliminary findings. Results from these experiments were consistent with predictions and suggest that people hold—independently—both negative implicit attitudes toward groups and positive implicit attitudes toward individuals. Together, the results of these experiments suggest that, regardless of phenomenological character of the attitude, perceivers regard groups more negatively than individuals. Expectations of Competitive Outgroup Behavior The outgroup-schema hypothesis proposes that intergroup interactions are more competitive than interindividual interactions because participants expect other groups to be more competitive than other individuals. The third category of evidence consists of five studies that jointly indicate that such differential expectations indeed exist (Hoyle et al., 1989; Insko et al., 1993; Insko et al., 2001; Schopler et al., 1995; Schopler et al., 2001). In a follow-up to the above-described study of differential beliefs regarding other groups and other individuals, Hoyle et al. asked participants to rate the expected characteristics of an anticipated series of interactions using the previously developed agreeableness and abrasiveness scales. One independent variable related to whether participants expected to participate in the interaction as an individual or as a group member. A second independent variable related to whether participants expected to
THE PERCEPTION OF OUTGROUP THREAT
255
interact with an individual or with a group. The interactions were described to the participants as “everyday social situations” but they never actually occurred. Results revealed no significant effects for whether the partici pant anticipated interacting as an individual or as a group member. However, there was a significant main-effect for the anticipation of interacting with an individual or a group. Interactions with a group were anticipated to be more abrasive than were interactions with an individual. Although the effect for agreeableness was in the predicted direction (i.e., more anticipated agreeableness for interactions with an individual), it was not significant. Participants in the Hoyle et al. (1989) study were told that they would participate in an experiment concerning the ways in which people interact in everyday social situations. As described above, participants anticipated everyday social interactions with groups to be more abrasive than interactions with individuals, regardless of whether they anticipated acting as individual or as a group member. Are these differential expectancies of individual and collective others also evident in the PDG setting? Four studies are relevant to this question. In the previously described study testing—and confirming—the prediction that communication between individuals produces a greater increase in cooperation than does communication between groups, Insko et al. (1993) also assessed participants’ expectancies regarding the decision of their opponent. This experiment involved one trial with the three-choice PDG-Alt matrix, and participants’ expectations were assessed by asking them to distribute 100 percentage points among the three PDG-Alt alternatives according to the likelihood that the opponent had chosen or would choose X, Y, or Z. This assessment was taken either immediately after the practice trials (“before choice”) or immediately after the choice, but before participants received feedback regarding the opponent’s choice (“after-choice”). Note that because the “before-choice” assessment occurred before the manipulation of communication was introduced, no effect for the communication manipulation could have occurred in this condition. For our present purposes, we therefore focus on the results for the “after-choice” assessment of expected competition and expected cooperation. We proposed earlier that if there is a schema dictating distrust of an outgroup, communication between groups should be less credible and persuasive than communication between individuals. If this is true, communication between individuals should lead to a larger decrease in expected competitiveness and a larger increase in expected cooperation than communication between groups. Consistent with this prediction, analysis of the “after-choice” assessments of expected competitiveness indicated a significant individuals versus groups by communication versus no communication interaction. For individuals, communication decreased expected competitiveness significantly but for groups, this effect was nonsignificantly reversed. Analysis of the “after-choice” assessment of expected cooperation also indicated a significant individuals versus groups by communication versus no communication interaction. For both individuals and groups, communication increased expected cooperation significantly but the effect was more pronounced for individuals than for groups. The second experiment (Schopler et al., 1995) relevant to differential ex pectancies of individual and group opponents within the context of matrix games also involved one trial with the PDG-Alt matrix, and included an assessment of expectancies after participants made their choice but before they learned of the opponent’s choice. The results indicate that individuals expected less competition, more cooperation, and less withdrawal choices from their opponents than did groups. The third experiment (Schopler et al., 2001) involved one trial with the PDG matrix. Individuals expected more cooperation from their opponents than did groups. The fourth experiment (Insko et al., 2001) also involved one trial with the PDG matrix. In addition to a manipulation of interindividual versus intergroup interaction, this experiment included a manipulation of anticipated future interaction. Approximately half of the participants were told that they would interact for only one trial, whereas the other participants were told that they would interact for multiple trials.
256
THE PSYCHOLOGY OF GROUP PERCEPTION
Expectancies were again assessed after participants made their choice, but before they learned of the opponent’s choice. A significant individuals versus groups by single versus multiple trials interaction indicated that the greater distrust in intergroup relative to interindividual interactions was more pronounced in the single trial condition than in the multiple trials condition. However, the smaller difference between interindividual and intergroup interactions in the multiple trials condition remained significant. Anticipating continued interaction increased trust between groups but not to the level observed between individuals. Outgroup Fear Expressed in Intergroup Discussions and Retrospective Reasons If there is greater distrust between groups than between individuals, we can expect that discussions between groups will include more statements of distrust than will discussions between individuals. The fourth category of evidence is consistent with this expectation. Two studies (Insko et al., 1994; Schopler et al., 1995) examined the content of tape-recorded communication between individuals and groups. In the Insko et al. study, which involved two different n-person generalizations of the PDG, the tape-recorded discussions were coded for both explicit statements of distrust (e.g., “I don’t trust you”) and implicit statements of distrust (e.g., “Don’t cheat on this”). Compared to discussions between individuals, discussions between groups were characterized to a greater extent by statements of both implicit and explicit distrust. Because the Insko et al. study obtained similar results for implicit and explicit statements of distrust, the two types of statement were combined into a single index of distrust in the Schopler et al. study. This study again found that, compared to discussions between individuals, discussion between groups were characterized to a greater extent by statements of distrust. The discussion data indicate more expressed distrust between groups than between individuals. These findings are consistent with the results of three studies that assessed participants’ own written reasons for their decisions (Insko et al., 2001; Schopler et al., 1995, 2001). In the Schopler et al. (1995) study, coding of participants’ open-ended responses revealed that, compared to individuals, groups reported more retrospective distrust and less retrospective trust. Furthermore, retrospective distrust statements were significantly correlated with cooperative choices (−.48) and withdrawal choices (.50). Retrospective trust statements were significantly correlated with cooperative choices (.82) and competitive choices (−.58). Likewise, both the Insko et al. (2001) and the Schopler et al. (2001) studies found that groups reported more retrospective distrust than did individuals. Consistent with the idea that distrust, or fear, mediates the discontinuity effect, both studies demonstrated the interindividual-intergroup discontinuity effect was significantly reduced when distrust was held constant. In agreement with the schema-based distrust hypothesis, there is compelling evidence across studies that groups make more retrospective distrust statements than do individuals, and that these statements are associated with PDG choice behavior. Differential Memory for Interindividual and Intergroup Interactions Why do participants believe and expect relations between groups to be more competitive than relations between individuals? One possibility is that participants recall intergroup relations as more competitive than interindividual relations. The fifth category of evidence concerning the differential distrust of groups and individuals comes from three studies by Pemberton, Insko, and Schopler (1996) that investigated participants’ memory for intergroup and interindividual relations. The first two of these investigations required participants to recall and describe in brief phrases either intergroup interactions or interindividual interactions. Next, participants rated the interactions they described on the abrasiveness and agreeableness scales of Hoyle et al. (1989). Examples of recalled
THE PERCEPTION OF OUTGROUP THREAT
257
intergroup interactions were: “church groups at camp,” “UNC vs. Duke basketball game,” and “Pro-Life vs. Pro-Choice groups.” Examples of recalled interindividual interactions were: “Christine and I talking at the mall,” “Shannon and I camping last weekend,” and “stranger and I stuck in the elevator.” Consistent with the results of Hoyle et al., neither study found significant effects for type of recalled interaction on the items of the agreeableness scale. However, both experiments found significant effects for type of recalled interactions on all items of the abrasiveness scale, such that the intergroup interactions were rated as more abrasive than the interindividual interactions. In both studies these effects remained significant after interactions relating to sports and games were removed from the data set and therefore cannot be attributed to the tendency for competitive intramural sports to be between groups rather than between individuals. In the third study, participants were asked to recall interactions in one of four categories: cooperative interindividual, cooperative intergroup, competitive interindividual, and competitive intergroup. Thus, there were two independent variables: cooperative versus competitive interactions and interindividual versus intergroup interactions. The dependent variable was the number of recalled interactions. Analysis of the number of recalled interactions revealed two significant effects. First, participants recalled significantly more interindividual than intergroup interactions, presumably because the former occurred more frequently than the latter. Second, and of greater importance, there was an interaction indicating that the tendency for greater recall of interindividual than of intergroup interactions was more pronounced for positive interactions than for negative interactions. Follow-up tests of simple effects indicated that participants recalled more competitive than cooperative intergroup interactions, and marginally more cooperative than competitive interindividual interactions. The interaction effect remained significant after interactions relating to sports and games were removed from the data set. Differential Experience with Inter individual and Intergroup Interactions The above-described studies by Pemberton et al. (1996) demonstrated that interindividual interactions are recalled predominately as cooperative, whereas intergroup interactions are recalled predominately as competitive. One possible explanation for this finding is that people actually experience intergroup relations as more competitive than interindividual relations. Our sixth category of evidence is consistent with this possibility. It consists of two studies with a modified version of the Rochester Interaction Record (RIR; Reis & Wheeler, 1991) reported by Pemberton et al. (1996). In the first study, participants kept a record of their interactions over a one-week period. They were trained to distinguish among five types of interaction: one-on-one interactions, within-group interactions, one-on-group interactions, group-on-one interactions, and group-on-group interactions. One-on-group and group-on-one interactions were distinguished by whether the participant interacted with a group (one-ongroup) or whether a group in which the participant was a member interacted with an individual (group-onone). The meaning of the term group was conveyed to participants through examples. Participants classified their interactions and then rated them on separate competitiveness and cooperativeness scales. The prediction was for the three types of interaction involving groups (one-on-group, group-on-one, and groupon-group) to be rated as more competitive than the two types of interaction not involving groups (one-onone, and within-group). In light of past results for beliefs, expectancies, and memories (e.g., Hoyle et al., 1989), no differences were expected among the five types of interaction on the cooperativeness scale. Consistent with predictions, the mean rated competitiveness of the three types of interaction involving groups was significantly higher than the mean rated competitiveness of the two types of interaction not involving groups. This difference remained significant after all interactions involving sports or games were eliminated. There were no significant differences among the different types of interaction for rated
258
THE PSYCHOLOGY OF GROUP PERCEPTION
cooperativeness. These results suggest that a possible explanation for the tendency to recall intergroup relations as more competitive than interindividual relations is that intergroup relations are actually experienced as relatively more competitive. This brings us to an important point: Schemas involve some abstraction away from the experiences from which they originate (Anderson, 2000; Smith, 1990). Thus, evidence for the existence of a schema is most obviously provided by results that cannot be explained simply by differential experience. In their first RIR study, Pemberton et al. (1996) addressed this issue directly by asking participants, after the week of recording their interactions, to recall as many of their one-on-one, group-on-one, one-on-group, and groupon-group interactions as possible. Subsequently, participants rated the recalled interactions for competitiveness. Mean competitiveness ratings on a 7-point scale for recalled interactions were 2.92 for one-on-one and 4.59 across the three types of interaction involving groups. For recorded interactions, the competitiveness ratings were 2.34 for one-on-one and 3.45 across interactions involving groups. The difference between one-on-one and interactions involving groups was significantly larger for the recalled interactions than for the recorded interactions, suggesting that the memory results were not just a function of experience. The second RIR experiment used a different procedure. Analogous to Pemberton et al.’s (1996) third memory study described earlier, participants did not rate their interactions but instead categorized each of the five types of interaction as either competitive or cooperative. The dependent variable was the number of recorded interactions in each of the resulting 10 categories over a one-week period. Conceptually replicating the findings of the first RIR experiment, results for the second RIR experiment indicated that the three types of interaction involving groups (one-on-group, group-on-one, and group-on-group) were seen as more competitive than the two types of interaction not involving groups (one-on-one and within-group). Results remained unchanged after all recorded interactions relating to sports and games were removed from the data set. Why are intergroup interactions experienced as more competitive than interindividual interactions? There are several possibilities. First, as some of the research reviewed thus far suggests, the beliefs and expectations associated with the outgroup schema can be self-fulfilling. When distrust and suspicion of other groups leads people to approach those groups with some degree of hostility, the other group may have no other option than to reciprocate this hostility, thus confirming the negative beliefs and expectations. Second, as suggested by the social support explanation of the discontinuity effect, intergroup interactions may be more competitive because group members can provide each other with social support for the pursuit of immediate self-interest. Third, as suggested by the identifiability explanation of the discontinuity effect, intergroup interactions may be more competitive because group membership provides a shield of anonymity. Fourth, as suggested by Schopler et al. (2001), interindividual relations, relative to intergroup relations, may afford greater freedom of choice when it comes to selecting interaction partners. The obvious advantage of having such freedom of choice is that one can choose to interact with cooperative partners and shun competitive partners. Section Summary We defined a schema as a set of beliefs serving to organize and guide memory for past events and expectations regarding future events. We proposed that the interindividual-intergroup discontinuity effect is partially rooted in a negative schema of outgroups, consisting of learned beliefs or expectations that intergroup interactions are competitive, unfriendly, deceitful, and aggressive. Six categories of evidence indicated that (1) participants make PDG-Alt choices indicative of more distrust of groups than of
THE PERCEPTION OF OUTGROUP THREAT
259
individuals; (2) participants believe that intergroup interactions are more negative than interindividual interactions; (3) participants expect other groups to act more competitively than other individuals; (4) participants express more distrust of other groups than of other individuals in discussioon and written statements; (5) participants remember interactions involving groups as more competitive than interactions not involving groups; and (6) differences in memory for interactions involving groups and interactions not involving groups in part reflect actual differences in experience. Results from the first RIR experiment of Pemberton et al. (1996) suggest, however, that the tendency to remember intergroup interactions as more competitive than interindividual interactions is not purely a function of experience. THE OUTGROUP SCHEMA: ACTIVATION We propose that the beliefs and expectations associated with the outgroup schema are activated when an aggregate of individuals is perceived as a group, or “social entity” (Campbell, 1958). In this section we review research findings that are consistent with the possibility that procedural interdependence among outgroup members is an antecedent of outgroup schema activation. Procedural interdependence refers to an interrelationship of group members’ behavior and outcomes. It is created, for instance, by introducing a consensus requirement or majority rule. Following such procedures, group members’ individual pref erences or decisions are combined into a collective group decision. We suggest that procedural interdependence among outgroup members activates the outgroup schema because it gives rise to the perception of entitativity or “groupness” (Campbell, 1958). Although this part of our chapter focuses on the role of procedural interdependence, we do not mean to imply that procedural interdependence is a necessary antecedent of outgroup schema activation. It is possible that other variables that increase the perceived entitativity of outgroups (e.g., spatial proximity) can lead to outgroup schema activation. The Role of a Consensus Rule Two early studies (Insko et al., 1987; 1988) demonstrated that a consensus rule (i.e., a requirement within both interacting groups for the members to reach a consensus on their PDG choice) contributes to the interindividual-intergroup discontinuity effect. To the extent that the discontinuity effect is driven by greater fear of other groups than of other individuals, these studies suggest that a consensus rule plays a role in outgroup schema activation. Note, however, that neither study included assessments of expectations regarding the opponents decision or participants’ retrospective reasons for their own decision. The evidence from these studies regarding the role of a consensus rule in outgroup schema activation is therefore suggestive rather than conclusive. The Insko et al. (1987) study was designed to test the altruistic-rationalization explanation of the discontinuity effect. According to this explanation, which has since been abandoned, the discontinuity effect arises from the tendency of group members to rationalize their competitive behavior as being an altruistic act, carried out for the sake of other group members. The study was run in a laboratory suite containing a center room with three smaller rooms on opposite sides and used a one-way experimental design with four conditions. The first condition was an individuals condition, in which three individuals, located in separate rooms on one side of the suite, interacted on an interindividual basis with single individuals located in separate rooms on the other side of the suite. On a given trial, participants examined the PDG matrix for 20 seconds, met in the center room for 15 seconds to discuss the situation with their opponent, and then returned to their homerooms where they had 15 seconds to select the cooperative or competitive choice. This procedure was repeated for 10 trials. The outcome interdependence condition was similar to the
260
THE PSYCHOLOGY OF GROUP PERCEPTION
individuals condition but in addition participants were told that they would be sharing outcomes with the two other persons on the same side of the suite. This created the opportunity to engage in altruistic rationalization. In the group-rep condition participants on the same side of the suite were placed in the same homeroom. Each group of three participants selected a representative to meet with a representative from the other group in the center room. After the representatives returned to their homerooms, both groups made their choice following a con sensus rule. The group-all condition was identical to the consensus condition, except that the members of both groups were required to meet collectively in the center room. Relevant means are presented in the top row of Table 18.1. Please note that the group-rep condition is labeled consensus in Table 18.1 in order to emphasize its resemblance to similar conditions in experiments to be described below. The group-rep and group-all conditions were more competitive than the individuals and outcome interdependence conditions. Inconsistent with the altruistic rationalization hypothesis, the latter two conditions did not differ significantly. In a follow-up experiment, Insko et al. (1988) examined what makes the group-rep and group-all conditions so much more competitive than the individuals and outcome interdependence conditions. Like the Insko et al. (1987) study, this study used a one-way experimental design, starting with an outcome interdependence condition, which was followed by four other conditions that progressively took on more of the properties of the group-all condition. In the second, or contact, condition the three participants on each side of the laboratory were placed in the same room but were not allowed to talk to each other. In the third, or discussion, condition the three participants in each room were required to discuss their individual PDG decision with each other. The fourth, or consensus, condition was one in which participants were required to reach a consensus regarding their separate PDG choices within their group. The fifth condition, labeled group-all, involved collective interaction between members of both groups, similar to the Insko et al. (1987) study. The only aspect in which the latter two conditions differed was that in the consensus condition the participants on one side of the laboratory still interacted individually with a member of the other group, whereas in the group-all condition, all three participants on one side of the laboratory interacted collectively with all participants on the other side. TABLE 18.1. Proportion of Competitive Choices Across Trials for the Insko et al. (1987), Insko et al. (1988), and Wildschut et al. (2001) Experiments Study
Individuals
Interdependenc Contact e
Discussiona
Consensusb
Groupall
Insko et al. .07 .07 — — .53 .36 (1987) Insko et al. — .15 .13 .17 .47 .50 (1988) Wildschut et al. (2001) U.S. sample .07 — — .32 .43 — Dutch sample .24 — — .30 .53 — Weighted .12 .10 .13 .26 .49 .42 Average (by N) Note. a This condition was labeled procedural independence by Wildschut et al. (2001); b This condition was labeled group-rep by Insko et al. (1987), and procedural interdependence by Wildschut et al.
THE PERCEPTION OF OUTGROUP THREAT
261
The second row of Table 18.1 shows that participants in the consensus condition were more competitive than those in the outcome interdependence, contact, and discussion conditions and did not differ significantly from those in the group-all condition. These results suggest that a consensus rule is a sufficient antecedent of the discontinuity effect. Unfortunately, these two studies are limited because expectations regarding the opponent’s decision or participants’ retrospective reasons for their own decision were not assessed. A third study (Wildschut, Lodewijkx, & Insko, 2001) did assess participants’ retrospective reasons and provides stronger support for the role of a consensus rule in outgroup schema activation. Dutch and US participants interacted with a cooperative programmed opponent on three trials of a PDG. In addition to participant nationality, a second independent variable related to the type of interaction participants were involved in. This independent variable had three levels. The first level was an individuals condition, in which participants were under the impression that they exchanged decisions with another individual through an audiovisual circuit. The second level resembled the discussion condition of Insko et al. (1988). Groups of two or three participants were under the impression that they exchanged decisions on an interindividual basis with members of another group through an audiovisual circuit. Although each group member interacted individually with a person in the other group, members of the same group were allowed to discuss their PDG decisions. The third level resembled the consensus condition of Insko et al. In this condition, participants were required to reach a collective agreement regarding their group choice. Upon completion of the three PDG trials, participants completed a questionnaire that included two items assessing fear of the opponent (“I do not trust the other person/group” and “I want to defend myself/ my group against the actions of the other person/ group”). The remaining items assessed two components of greed: a concern for maximizing own absolute outcomes, or max own, and a concern for maximizing own outcomes relative to the outcomes of the opponent, or max rel. Mean proportions of competitive choices for the US and Dutch samples are presented in the third and fourth row of Table 18.1, respectively. Across nationalities, participants in the consensus condition were more competitive than were participants in the individuals and discussion conditions. The latter two conditions did not differ significantly. There were no significant effects involving nationality. The results for retrospective fear tracked the choice data. Retrospective fear was higher in the consensus condition than in the discussion and individuals conditions, which did not differ significantly. When fear was controlled, the main effect for interaction type (i.e., individuals vs. discussion vs. consensus) was no longer significant. In this model, the association between fear and competition was significant. This pattern of results is consistent with the possibility that outgroup fear mediates the role of a consensus rule in interindividualintergroup discontinuity. However, the fact that fear was assessed after the PDG interaction means that the association between fear and choice behavior may reflect the effect of choice behavior on fear, instead of the reverse. Jointly, these three studies provide compelling evidence for the role of a consensus rule in interindividualintergroup discontinuity. To the extent that intergroup competition is driven by outgroup fear, results from the first two studies are consistent with the possibility that a consensus rule leads to activation of the outgroup schema. Mediation analyses conducted within the context of the third study support this conclusion. Outgroup Fear in the Absence of a Consensus Rule Is a consensus rule a necessary antecedent of discontinuity effect? This question was addressed in a study by Insko et al. (1994; Experiment 2). In this experiment, individuals, groups without required consensus,
262
THE PSYCHOLOGY OF GROUP PERCEPTION
and groups with required consensus interacted within the context of one of two n-person generalizations of the PDG: the intergroup public goods game (IPG) and the intergroup prisoner’s dilemma (IPD) game. The IPG is based on the minimal contributing set paradigm and models intergroup conflict over step-level public goods (Rapoport & Bornstein, 1987). The IPD game models intergroup conflict over continuous public goods (Bornstein, 1992). In the IPG and IPD, players (individuals or groups) are given promissory notes that can be invested. Self-interest is maximized by investing more notes than the opponent. Similar to the PDG, however, when both players invest all their notes, they earn less than they would have earned if neither player had invested their notes. In other words, investing in the context of these n-person generalizations of the PDG is analogous to competing in the context of a two-choice PDG. Results indicated that groups were more likely to invest (i.e., were more competitive) than individuals in both the IPG and IPD. Seemingly inconsistent with the experiments of Insko et al. (1987, 1988) and Wildschut et al. (2001), there was no significant effect for required consensus on intergroup competition. These results demonstrate that the interindividual-intergroup discontinuity effect can occur in the absence of a consensus rule. A number of additional dependent variables were derived by coding tape-recorded discussions for various types of statement, including explicit fear and implicit fear. Participants in the individual condition made fewer explicit and implicit fear statements than did participants in the two group conditions. The norequired-consensus condition and required-consensus condition did not differ on explicit fear or implicit fear. Thus, greater fear of other groups than of other individuals can occur in the absence of a consensus rule. Procedural Interdependence and Outgroup Fear Three studies (Insko et al., 1987, 1988; Wildschut et al., 2000) demonstrated the role of a consensus requirement in interindividual-intergroup discontinu ity. Findings from the most recent of these studies suggest that the effect of a consensus rule is mediated by outgroup fear. In light of these findings, it seems surprising that within the context of two n-person generalizations of the PDG no significant effects were found for a consensus requirement on either intergroup competition or statements reflecting outgroup fear. Note, however, that consensus decisions in the two-choice PDG and group decisions in the n-person generalizations of the PDG share a key feature. In each case, the group members’ decisions are translated into a collective group choice. In the case of consensus decision with the two-choice PDG this is straightforward: the group members discuss their preferences during a within-group discussion period and decide on the collective group choice. Within the context of the n-person generalizations of the PDG this is more complex. Although group members decide individually whether or not to invest their promissory note, the total number of notes invested by the group as a whole is determined by the joint actions of the individual group members. In the n-person games, the total number of notes invested by one group is combined with the total number of notes invested by the other group to determine each group’s total payoff. Thus, in both the two-choice PDG and the n-person games there is an interrelationship of own group members’ decisions and outcomes, a state of affairs that has been labeled procedural interdependence (Insko et al., 1994, Wildschut et al., 2001). With the two-choice PDG, a consensus rule may be a prerequisite for procedural interdependence whereas with the more complex n-person games a consensus rule may not be required. An important remaining issue is that in these three studies the effects of ingroup and outgroup procedural interdependence were confounded. Within each experimental session, members of both groups either were or were not procedurally interdependent. The observed effects of procedural interdependence on
THE PERCEPTION OF OUTGROUP THREAT
263
competition and outgroup fear could therefore be due to procedural interdependence among members of the ingroup, procedural interdependence among members of the outgroup, or both. We propose that outgroup procedural interdependence, but not ingroup procedural interdependence, is an antecedent of outgroup schema activation. Ingroup procedural interdependence is, in our view, more obviously relevant to the identifiability and social support explanations of the discontinuity effect. Is the exclusive focus on outgroup procedural interdependence as an antecedent of outgroup schema activation justified? Recall Hoyle et al.’s (1989) second experiment in which participants rated the expected characteristics of an anticipated series of interactions for agreeableness and abrasiveness. One independent variable related to whether participants expected to participate in the interaction as an individual or as a group member. A second independent variable related to whether participants anticipated interacting with an individual or with a group. Consistent with the postulated outgroup schema, interactions with a group were expected to be more abrasive than were interactions with an individual. More important, there was no significant effect for whether participants anticipated acting as individual or as group member. In other words, target attributes were more important in determining ratings of abrasiveness than perceiver attributes. These findings partially justify our focus on procedural interdependence of the outgroup rather than procedural interdependence of the ingroup. A more compelling justification for our focus on outgroup procedural interdependence comes from a recent experiment by Insko, Wildschut, and Pinter (2003), which involved one trial with the PDG matrix. One independent variable related to whether or not outgroup members were procedurally interdependent. A second independent variable related to whether or not ingroup members were procedurally interdependent. An effect for outgroup procedural interdependence on intergroup competition would be consistent with the hypothesized link between outgroup procedural interdependence and outgroup schema activation. The role of ingroup procedural interdependence is not directly relevant to this hypothesized link, although this is not to say that an effect for ingroup procedural interdependence would be uninteresting or unimportant. Participants were run in groups of three and were seated individually in three small rooms that were attached to a larger central room. Participants were told that they would be interacting with another group of three persons, who were seated in a nearby laboratory. In reality, no other group was present and all feedback from this alleged other group was controlled by the experimenter. Participants were told that they were yoked to a person in the other group and that they would interact with this person in the context of a PDG. Procedural interdependence among ingroup members was manipulated by instructing participants to use one of two different decision rules. In the no-procedural interdependence condition, participants were told that they would determine individually whether to compete or cooperate with the outgroup member to whom they were yoked. In the procedural interdependence condition, participants were told that there would be a group decision, and that this decision would be determined following a majority rule. In this condition, participants could not determine individually whether to compete or cooperate with the outgroup member to whom they were yoked. Instead, their decision was determined by the majority of the group. Outgroup procedural interdependence was manipulated by informing participants that the outgroup members to whom they were yoked would either be following the individual or the majority decision rule described above. Consistent with the hypothesized link between outgroup procedural interdependence and outgroup schema activation, participants were significantly more competitive when they thought that outgroup members were following a majority decision rule than when they thought that outgroup members were following an individual decision rule. There was also an effect for ingroup procedural interdependence. Participants were more competitive when they thought that ingroup members (including themselves) were following a majority decision rule than when they thought that ingroup members were following an individual decision rule (for a full discussion of the latter finding, see Insko et al., 2003).
264
THE PSYCHOLOGY OF GROUP PERCEPTION
Mediation by Perceived Entitativity Why does outgroup procedural interdependence increase outgroup fear? Insko et al. (1994) proposed that procedural interdependence is an antecedent of perceived entitativity, or the perception of an aggregate of persons as a group (Campbell, 1958). Entitativity, in turn, may lead to outgroup schema activation. We propose this causal sequence from procedural interdependence to perceived entitativity to outgroup schema activation as suggestive rather than conclusive. Nevertheless, we believe that there is a strong theoretical basis for this proposed sequence, as well as some compelling empirical support. Campbell (1958) suggested that aggregates of persons are organized according to elementary perceptual principles identified by Gestalt psychologists. Specifically, Campbell suggested that the degree of similarity, proximity, and common fate among persons affects the perceived “groupness” of the aggregate. Common fate, which can be defined as the degree to which elements move in the same direction in space and time, shares some key features with procedural interdependence. By definition, procedural interdependence creates a situation where group members move in the same direction (i.e., the direction of the collective group decision) and experience common outcomes as a function of the group decision. S.L.Gaertner et al. (1999), referring to interaction between groups, noted that “when groups interact fully to reach a consensus solution to the same problem, members may intrinsically experience common fate with one another” (p. 390). In agreement with this idea, Insko et al. (1988) found that procedurally interdependent participants in the consensus and group-all conditions, who had to reach a collective decision within their group, reported less self-control over personal earnings than procedurally independent participants in the outcome interdependence, contact, and discussion conditions. The above-described research by Insko, Wildschut, and Pinter (2003) produced results consistent with the postulated causal sequence from outgroup procedural interdependence to outgroup entitativity to outgroup schema activation. Relevant to the first link in this sequence, they found that participants were more likely to perceive the interaction as involving two entitative groups as opposed to separate individuals when outgroup procedural interdependence was present rather than absent. Relevant to the second link in the sequence, they found that those participants who perceived the interaction as involving two entitative groups reported more distrust of the outgroup than those who perceived the interaction as involving separate individuals. Distrust, in turn, predicted competition. Further evidence for the seond link in the causal sequence was provided by Dasgupta, Banaji, and Abelson (1999). They found that greater perceptual similarity and spatial proximity among a group of person-like figures increased the perceived likelihood that the group would engage in negative behavior visà-vis another group of person-like figures. Assuming that similarity and proximity are antecedents of entitativity, these results can be interpreted as demonstrating that “participants see groups as more combative and dangerous when they are more entitive” (Abelson, Dasgupta, Park, & Banaji, 1998, p. 247). A final source of evidence is offered by research on perceptions of outgroup homogeneity. It has been demonstrated that perceptions of other groups become less negative when their heterogeneity is made salient (e.g., Maurer, Park, & Rothbart, 1995). Assuming that perceived heterogeneity of another group is inversely related to the perceived entitativity of this group, these results support the proposed effect of perceived outgroup entitativity on outgroup schema activation. Section Summary We reviewed research demonstrating that procedural interdependence is an antecedent of outgroup fear. These findings are consistent with the possibility that procedural interdependence leads to outgroup schema activation. We accounted for this relation between procedural interdependence and outgroup schema
THE PERCEPTION OF OUTGROUP THREAT
265
activation in terms of perceived entitativity. A causal sequence from procedural interdependence to perceived entitativity to outgroup schema activation was tentatively proposed. The first link in the sequence is in agreement with Campbell’s (1958) seminal analysis of group perception and is supported by recent research examining the unique effects of ingroup procedural interdependence and outgroup procedural interdependence on intergroup behavior and intergroup perception (Insko et al., 2003). The second link in the sequence is supported by research on the relation between antecedents of entitativity and outgroup fear (Abelson et al., 1998; Dasgupta et al., 1999). DISCUSSION The Reduction of Outgroup Fear Research on interindividual-intergroup discontinuity has attempted to identify strategies for reducing intergroup conflict by examining why intergroup relations are sometimes more hostile and competitive than interindividual relations. The schema-based distrust (or fear) hypothesis is based on the tendency to distrust other groups more than other individuals. According to this hypothesis, intergroup interactions tend to be more competitive than interindividual interactions because the anticipation of interacting with another group activates a negative outgroup schema, consisting of learned beliefs and expectations that intergroup interactions are competitive, deceitful, and aggressive. Because schemas involve some abstraction away from the experiences that give rise to them, evidence for the existence of a schema is most obviously provided by results that cannot be explained in terms of different experience. Recall that Pemberton et al.’s (1996) first RIR experiment provided such evidence. Participants were asked to recall as many of the interactions that they had recorded over the course of a week. Interactions involving groups were both recorded and recalled as being more competitive than one-on-one interactions, but this difference was significantly larger for the recalled interactions, suggesting that participants’ memory reflected some abstraction away from actual experience. If our beliefs and expectations of other groups are based on a summary representation abstracted from past experience, then outgroup fear may be difficult to reduce. Why is this the case? S.B.Klein and Loftus’s (1993) research on the representation of self-knowledge supports the idea that abstract self-knowledge and autobiographical self-knowledge are accessed independently and may in fact be stored independently. According to their Dual Exemplar/ Summary view, the self-descriptiveness of traits (e.g., friendliness) can be judged by accessing directly an abstract representation of the self (e.g., I am a friendly person), or by accessing trait-relevant autobiographical memories (e.g., I gave John a nice birthday present). Klein and Loftus propose that in the process of judging the self-descriptiveness of traits we rely primarily on a summary representation of the self and that autobiographical information is accessed only when a trait is not represented in summary form. To the extent that findings regarding the representation of self-knowledge can be generalized to the representation of knowledge about other groups— and there is evidence that supports such a generalization (Gramzow, L.Gaertner, & Sedikides, 2001; E.R.Smith, in press)—Klein and Loftus’s research implies that outgroup fear may be difficult to reduce. When people anticipate interacting with another group, they may rely on the summary information represented in the outgroup schema rather than their memory for specific interactions with particular outgroups, some of which may have been friendly. A possible strategy for reducing outgroup fear that is less susceptible to the above-described problem is suggested by the Common Ingroup Identity model of S.L.Gaertner, Dovidio, Anastasio, Bachman, & Rust (1993). According to this model, intergroup bias can be reduced “by transforming members’ cognitive
266
THE PSYCHOLOGY OF GROUP PERCEPTION
representations of the memberships from two separate groups to one involving a common ingroup identity” (S.L.Gaertner et al., 1999, p. 389). If this transformation is successful, so the model proposes, “processes that produce more favorable regard for ingroup members can be directed to former outgroup members” (S.L.Gaertner et al., p. 389). A further outcome of creating a common ingroup identity, or superordinate group (Sherif, Harvey, White, Hood, & Sherif, 1988), may be that processes producing unfavorable regard for outgroup members, such as outgroup schema activation, are eliminated. Past research on interindividual-intergroup discontinuity has revealed surprisingly little evidence for the role of group members’ cognitive representations in determining intergroup behavior. Insko et al. (1998) found that, compared to the usual situation where the PDG choices of both sides are unconstrained by the experimenter, interacting with an opponent who followed a tit-for-tat strategy reduced intergroup competition. This reduction in competition was not tracked, however, by group members’ cognitive representations of the interaction. Nevertheless, more recent studies have produced some interesting evidence for the relation between group members’ cognitive representations of the interaction and outgroup fear. For instance, Insko et al. (2001) asked participants to indicate whether they thought of the people participating in the PDG as one group, two groups, or separate individuals (S.L.Gaertner, Mann, Murrell, & Dovidio, 1989). Results revealed a significant association between the perceived likelihood that the other group would compete on the PDG and the two-groups representation, with the two-groups representation being stronger for groups that anticipated a competitive opponent than for groups that anticipated a cooperative opponent. This relation between trust and categorization suggests the interesting possibility that the tendency to perceive intergroup interactions as involving separate, entitative groups partially determines outgroup fear. Consistent with this possibility, participants in a minimal group experiment by L.Gaertner and Insko (2000) expected their monetary allocations (on Multiple Alternative Matrices) to be less strongly reciprocated by members of a different artistic-preference category than by members of their own artisticpreference category. Jointly, these studies indicate that one possible route to reducing outgroup fear is recategorization, or the creation of a common ingroup identity. Is Trust Enough? Recall that, in the introductory sections of this chapter, we proposed that the tendency for intergroup interactions to be more competitive than interindividual interactions is rooted in the greater fear and greed in intergroup as compared to interindividual relations. The question therefore remains whether possible reductions in outgroup fear flowing from interventions aimed at establishing a common ingroup identity will go hand-in-hand with reductions in greed. In other words, does trust provide a sufficient basis for cooperation between groups? Somewhat paradoxically, we believe that trust may sometimes increase cooperation and sometimes decrease cooperation (or increase competition). There are three possible reasons why trust might increase cooperation. First, trust carries the strong implication that in the long run outcomes can be maximized through cooperation. Second, trust contradicts a rationalization that is sometimes used to justify greed-based exploitation. Insko et al. (1993) cited anecdotal evidence that diplomats of the former Soviet Union sometimes justi fied Soviet domination of Eastern Europe on the basis of Russia’s historic fear of being invaded by the West; that is, by Soviet distrust of the West. Third, it is possible that trust carries with it the normative expectation of being cooperative. Taking advantage of a cooperative and vulnerable other may be counter normative—certainly if the other is an individual, but even if the other is a group. The outrage that Americans felt toward the Japanese following the Pearl Harbor attack may be partially explained on this basis.
THE PERCEPTION OF OUTGROUP THREAT
267
Given the above reasons why trust may increase cooperation, why should trust ever decrease cooperation? The reason is that in the context of a PDG-like situation trust implies that the other will be cooperative and thus vulnerable to exploitation (Axelrod, 1984; Shure, Meeker, & Hansford, 1965). This, of course, assumes that individuals or groups are sometimes likely to respond on the basis of simple selfinterest. It seems apparent that in some situations individuals, and particularly groups, do respond on the basis of own interest. What might contradict this tendency of trust to increase competition? This is an interesting question for future research, but one possibility relates to perceiving common categorization in the context of continuing interaction, where self interest can be maximized through long-term cooperation as opposed to short-term competition. AUTHOR NOTE Preparation of this chapter was supported by NIMH grant MH 53258 to Chester A. Insko. We wish to thank Olivier Corneille, Jennifer Dent, Ana Guinote, Dick Moreland, and Constantine Sedikides for their helpful comments on previous versions of this chapter. Correspondence concerning this chapter should be addressed to Tim Wildschut, School of Psychology, University of Southampton, Southampton SO17 1BJ, England, UK. Electronic mail may be sent to
[email protected].
19 The Maintenance of Entitativity A Subjective Group Dynamics Approach DOMINIC ABRAMS University of Kent JOSÉ M.MARQUES University of Porto GEORGINA RANDSLEY DE MOURA PAUL HUTCHISON University of Kent NICOLA J.BOWN Leeds University Business School
Groups have a particular quality that fosters a desire for consensus and compliance (e.g., Asch, 1952; Boyanowsky & Allen, 1973; Burnstein & Vinokur, 1975; Festinger, 1950; Hogg & Hains, 1998; Insko, Drenan, Solomon, Smith, & Wade, 1983; Janis, 1982; Levine, 1989; Sherif, 1936). Disunity potentially undermines both the social reality offered by the group and the potential to attain group goals, so it is not surprising that group members conform themselves and make attempts, often successfully, to induce conformity to the group (see Berkowitz & Howard, 1959; Davis & Witte, 1996; Deutsch & Gerard, 1955; Earle, 1986; Latané & Nida, 1981; Levine & Thompson, 1996; Mullen, 1991; Schachter, 1951; Williams, Cheung & Choi, 2000). As M.E.Shaw (1976) observed, “An individual who can be expected to behave in conventional ways is unlikely to disrupt the group, whereas an unconventional person is likely to cause disorder and dissatisfaction.” (p. 176). This perspective on group processes suggests that reactions to group members will be driven substantially by the magnitude of their differences from the group norm. Contrary to the traditional view of small group processes summarized above, an important feature of the social identity approach is that groups and intergroup relationships may have their impact on perception and behavior entirely through the process of categorization. Social Identity Theory (SIT) holds that when social identity is salient, a person attributes category-based features to all category members, thereby minimizing individual differences within categories, and maximizing intercategory differences (e.g., Abrams & Hogg, 1990; Hogg & Abrams, 1988; Tajfel, 1974; Tajfel & Turner, 1979; see also Brewer, 1988; Fiske & Neuberg, 1990). Self-categorization theory (SCT) strengthens this idea. SCT considers two aspects of the fit between individuals and social categories, comparative and normative fit (Oakes, 1996; Oakes, Turner, & Haslam, 1991). According to this theory, perceptions of group members are determined by a meta-contrast, which can be approximated mathematically as a ratio of intragroup differences versus intergroup differences (Hogg & McGarty, 1990; Oakes, Haslam, & Turner, 1994; Turner, Hogg, Oakes, Reicher, & Wetherell, 1987). This contrast produces abstract prototypes that represent the positions (e.g., on an attitude continuum)
THE MAINTENANCE OF ENTITATIVITY
269
that best capture differences between the ingroup and outgroup to the detriment of intracategorical differences. Deviant category members would thus tend to be overlooked or disregarded under the operation of the metacontrast principle, particularly when the categories are highly salient. If the deviant is not, or cannot be, disregarded or recategorized, the metacontrast is reduced. Therefore, one consequence could be that perceivers should use a more relevant categorical dimension. An alternative cognitive response would be to assimilate the group prototype toward the position held by the deviant, a process that would likely depend on the extremity of deviance (e.g., Kunda & Oleson, 1997). This would maximize the fit between the social categorization in use and the characteristics of the people being categorized (Oakes et al., 1991, 1994). Both consequences would be psychologically satisfying to the extent they would clarify intergroup boundaries (see Hogg, 1993). Social cognition research on entitativity has recently revived Campbell’s (1958) idea that the entitativity of a group is based primarily on four defining features: proximity, similarity, common fate and collective movement (see Abelson, Dasgupta, Park, & Banaji, 1998; Sherman, Hamilton, & Lewis, 1999). Other features, such as prototypicality, essentialism, group size, intragroup interdependence, common goals, common outcomes, and importance of group membership, also seem to contribute to entitativity to some extent (Brewer & Harasty, 1996; Gaertner & Schopler, 1998; Lickel, Hamilton, Wieczorkowska, Lewis, Sherman, & Uhles, 2000; McGarty, Haslam, Hutchinson, & Grace, 1995; Mullen, 1991; Sherman et al., 1999; Yzerbyt, Rocher, & Schadron, 1997). Consistent with the idea that entitativity is a desired property of groups, Lickel et al. (2000) found that participants attached greater value to membership in more entitative groups, as shown by higher correlation between ratings of the importance of membership in those groups and perceptions of entitativity. Research on entitativity further suggests that the presence of deviant group members should be especially problematic for more normative members because it highlights potential disunity, and should thereby reduce entitativity. Consistent with this idea, Sherman et al. (1999) proposed that “members of highly entitative groups will perceive greater differentiation from outgroups and thus show a greater degree of ingroup bias in perceptions and interpretations of events… [In addition] entitative ingroups should be seen as having more power to do good things and to achieve positive goals…highly entitative groups are more likely to develop clear group norms” (p. 102). Thus, the traditional small group approach, the social identity approach, and the social cognition approach are all consistent in arguing that there is a positive relationship between attachment to the group and higher ingroup uniformity, entitativity, or distinctiveness. The presence of deviant members could potentially imperil group entitativity. The present chapter explores a motivated cognition model that examines how group members may respond to threats to, and exert influence over, important aspects of group entitativity. SUBJECTIVE GROUP DYNAMICS The Subjective Group Dynamics model (SGD; Marques, Páez, & Abrams, 1998; Marques, Abrams, Páez, & Hogg, 2001) proposes that people generally strive to confirm ingroup reality. For this to happen, at least two conditions have to be met. First, the ingroup must be entitative (Campbell, 1958). In intergroup contexts this requires both intergroup distinctiveness and intragroup uniformity. Second, there must be social validation of the ingroup perspective as opposed to the outgroup perspective. For this to be achieved, group members must differentiate other members who strengthen from those who weaken such validation. In short, we argue that group members are motivated to maximize and sustain descriptive intergroup distinctiveness whilst maximizing and sustaining the relative validity of prescriptive ingroup norms through the parallel and complementary process of intragroup differentiation (Marques, Abrams, Páez, & Martinez-Taboada, 1998).
270
THE PSYCHOLOGY OF GROUP PERCEPTION
Subjective Group Dynamics and Deviance The SGD model proposes that judgments about deviant group members depend on the interplay between self-categorization, denotative and prescriptive norms, and social self-regulation. In line with SCT, we believe that the categorization process is largely driven by a search for meaning and reduction of uncertainty (e.g., Hogg, 2000a). It follows that people are motivated to ensure the validity of their subjective sense of reality, provided by the reality defined by the ingroup (Abrams, 1990, 1992; Abrams & Hogg, 1988, 2001; Hogg, 2001b; Marques & Páez, 1994). This certainty is strengthened to the extent that self and ingroup are seen as sharing a common set of norms and values (e.g., Turner, 1991; see also Cadinu & Rothbart, 1996; Krueger & Clement, 1996). The flexibility of self-categorization allows mismatches between individuals and groups to be avoided or resolved by the application of different categorizations reflecting the intergroup context (see Abrams, 1996, 1999a; Spears, Oakes, Ellemers, & Haslam, 1997; Turner & Oakes, 1997). However, in some situations a specific group member who cannot be recategorized may diverge from normative expectations. A deviant member such as this is problematic because he or she undermines the consensus of the group and therefore undermines the subjective validity of the social reality. The SGD model holds that judgments and evaluations of deviant group members reflect two interrelated goals. As predicted by SIT, group members wish to ensure that ingroups have higher value than relevant outgroups. In addition, in line with SCT, group members wish to maintain the validity of ingroup norms and thereby validate their subjective reality. We believe that ingroup superiority is often achieved through global category differentiation, whereby the ingroup is favored over the outgroup (see Mullen, Brown, & Smith, 1992). However, validation of ingroup norms often depends on making distinctions within groups to determine which individual members either reinforce or undermine those norms. We assume that judgments of individual group members remain essentially depersonalized; that is, they are framed with reference to group norms and stereotypes. The intergroup and intragroup processes operate in conjunction in such a way that it becomes possible, rather than paradoxical, that someone who strongly favors the ingroup over the outgroup may also prefer some outgroup members over some ingroup members. Indeed, we suggest that groups and individuals are judged in a way that is functionally complementary. Denotative Norms To develop this idea we distinguished between denotative and prescriptive norms (Marques, Abrams, Páez, & Martinez-Toboada, 1998; Marques et al., 2001; Marques et al., 1998). Denotative norms provide the descriptive criteria for categorization and are thus relevant to the meta-contrast principle and both comparative and normative fit, as defined in SCT. Denotative norms are perceived as essential for, inherent in, or entirely indicative of category membership. Therefore compliance with the norm and category membership are likely to be perceived as inextricable, and category ascriptions can be made immediately from a person’s adherence to these norms. For example, hearing a deep voice leads to a strong inference that the speaker is male. In turn, ingroup bias may result from a category membership inference based on denotative differences. It seems likely that denotative norms are often applied nonconsciously or at least relatively unreflectively. There are also occasions when people do devote conscious attention to denotative norms, for example when there is a high degree of ambiguity regarding category memberships (e.g., Abrams, 1990, 1996; Abrams & Brown, 1989; Abrams & Masser, 1998), when it is important to ensure that no outgroup members are categorized as ingroup members (Yzerbyt, Castano, Leyens, & Paladino, 2000), when perceivers are prejudiced (Blascovich, Wyer, Swart, & Kibler, 1997), or when they need to preserve cognitive closure (Kruglanski & Webster, 1996). Norms that denote category membership may be strongly
THE MAINTENANCE OF ENTITATIVITY
271
associated with preferential judgments (e.g. sports fans may favor others who wear the same colors as themselves), but they are not the sole basis for evaluation of ingroup and outgroup members. Prescriptive Norms Sometimes it is clear that ingroup and outgroup members should behave differently to uphold their differences (e.g., the expectation that Jews should support Israel, Arabs should support the PLO). When groups hold competing goals or viewpoints, these oppositional norms are usually very clear for members of both groups. In addition, there are norms and standards which are not oppositional but which are still very important for ingroup members (see Forsyth, 1990). For example, ingroups may desire to embody societal, or generic, norms to a greater extent than outgroups. Both oppositional and generic norms can take on a prescriptive character. We believe that, as well as distinguishing people’s group memberships based on denotative (or descriptive) differences between categories, group members also distinguish individual group members from one another in terms of how far they support the prescriptive norms of the ingroup. Indeed, these two processes should work in conjunction. Reactions to a particular group member will depend on whether he or she appears to be an ingroup or outgroup member and whether his or her behavior appears to undermine or validate the ingroup prescriptive norm. For example, imagine someone commits a faux pas at a social gathering. To the extent that the ingroup-outgroup distinction is salient, it is likely to be more discomforting to learn that the person is an ingroup member than an outgroup member because embarrassment for the ingroup is equally embarrassing for the self. We propose that evaluations of specific group members depend on a focus on prescriptive norms and that this often involves backward processing (see Marques et al., 2001), a form of counterfactual thinking that occurs when observed events run counter to expectations (D.T.Miller & Prentice, 1996). In these situations, people generate a specific frame of reference that accounts for the counterintuitive event, and they construct, online, a standard of comparison relevant to that particular context (Kahneman & Miller, 1986). Deviants violate normative expectancies, and this makes prescriptive norms highly salient as standards against which to judge ingroup and outgroup behavior. These judgments reflect the evaluative consequences that group members’ characteristics and behavior have for the ingroup and hence for the social self. Ingroup members attend to prescriptive norms so as to ensure consensus on criteria for positive ingroup evaluation. Salient variations from prescriptive norms are therefore very likely to induce active regulation of the subjective image of the group. Specifically, because group members are motivated to preserve the subjective validity of their group’s norms, they will wish to correct or remove challenges to those norms within the group and to gather evidence from outside the group that bolsters ingroup norms. Prescriptive norm differentiation is crucial because it provides a means of garnering evidence that consolidates the validity of ingroup reality. Prescriptive norm differentiation serves a function that cannot be met by categorization alone, because it allows group members to confirm their hypotheses, not about differences between groups, but about the value they associate with their own group membership. For example, the value of ingroup consensus is often made all the more real when a member breaks ranks or deviates from the group norms (e.g., Holtz & Miller, 1985; N.Miller, Gross, & Holtz, 1991). In the McCarthy era, the threat to the United States from Americans who might support Communist ideology was used as a basis for justifying and intensifying the cold war. The issue was not whether people were American or Russian, but whether they were engaging in so-called un-American activities. Category membership was not in doubt; what mattered was violation of ingroup prescriptive norms.
272
THE PSYCHOLOGY OF GROUP PERCEPTION
Subjective Group Dynamics and Social Self-Regulation Prescriptive norm differentiation within groups is a process that allows the relationship between different social categories to become dynamic. Believing that the ingroup is somehow better or has more valid claim to spirituality than the outgroup legitimizes category-based favoritism (based on denotative differences). Once group members are aware of differences of opinion or behavior within the group, they have a motive to validate this subjective reality by enhancing ingroup consensus. Deviant members may shake the subjective reality of the group, and therefore they require special attention. Sensitivity to deviation from prescriptive norms is necessary so that group members can respond to sustain the subjective validity of the ingroup’s position. Carver and Scheier (1981, 1998) proposed a theory of self-regulation based on a cybernetic framework in which higher order principles (e.g., to maintain a tidy house) set the standards for specific goals (e.g., to throw out the piles of old newspapers). Whenever there is a mismatch between standards and attainments at one level of control, the system intervenes to ensure this is corrected at the immediately subordinate level. When routine operation of the system is dis rupted (e.g., when behavior does not match the relevant standard), attention is devoted to reducing discrepancies between the intended and actual state (see also Bodenhausen & Macrae, 1998; E.T.Higgins, 1987). The Social Self-Regulation (SSR) model proposed by Abrams (1990, 1994, 1996, 1999a) extends selfregulation theory to apply to the domain of social identity. Within intergroup contexts, the overall goals of behavior are predicted primarily from the relative salience of personal versus social identity (Abrams, 1999a; Abrams & Masser, 1998). However, the specific choice of behavior may vary depending on how the goals can most easily or appropriately be achieved. Therefore, when adherence to a group goal is disrupted, group members may engage in corrective behavior that diverges from routine forms of intergroup differentiation and intragroup conformity. The SSR model holds that novel intergroup situations, conflicting role demands, or disruptions to group standards are all likely to require group members to stop and think— to select actions consciously and strategically. Different specific goals or standards can arise from a variety of factors. These include the nature of the intergroup context (e.g., competitive vs. cooperative intergroup relations), group members’ motivation to sustain a positive identity, their skills and ability to enact certain behaviors, and anticipated responses from a potential audience. There is a variety of direct and indirect evidence that self-regulation processes are engaged to influence intergroup and intragroup behavior (e.g., Abrams, 1985; Abrams & Brown, 1989; Bodenhausen & Macrae, 1998; Monteith, Sherman, & Devine, 1998; Plant & Devine, 1998; see also Reicher, Spears, & Postmes, 1995; Spears, 2001). We believe that social self-regulation processes also underpin reactions to ingroup and outgroup deviance because, when social identity is salient, maintenance of ingroup standards is a means of validating the standards that are used to regulate the self. Therefore, when social identity is salient or important, one aspect of group members’ self-regulation is the regulation of group adherence to group standards. The presence of a deviant group member indicates that the group is failing to sustain its norms and values. Judgments of, and reactions to, a deviant will therefore depend on whether the target person is an ingroup or an outgroup member and also whether the deviant’s behavior is more or less consistent with the ingroup’s values or norms. For the perceiver the important issue is whether the group member’s behavior provides a source of validation for ingroup norms, either directly or by undermining outgroup norms in relative terms.
THE MAINTENANCE OF ENTITATIVITY
273
THE BLACK SHEEP EFFECT: A CASE FOR SUBJECTIVE GROUP DYNAMICS One important source of evidence for the SGD model is the fact that people tend to evaluatively upgrade attractive ingroup members and downgrade unattractive ingroup members, as compared to analogous outgroup members (for a review, see Marques & Paez, 1994). This phenomenon was coined the ‘black sheep effect’ (Marques, Yzerbyt, & Leyens, 1988). Evidence shows that the black sheep effect occurs even when individuals show a strong overall preference for the ingroup (Marques, Robalo, & Rocha, 1992). As long as the intergroup context is salient, the effect is larger when individuals identify with the ingroup (Branscombe, Wann, Noel, & Coleman, 1993), and can occur when perceivers judge members singly (Marques et al., 1988), or directly compare normative and deviant members from the same group, or two members from different groups (Marques & Yzerbyt, 1988). Evidence also suggests that the black sheep effect emerges when differences between group members are relevant to the maintenance of positive ingroup valence (Marques, 1990) or to intergroup distinctiveness (Marques et al., 1988, Experiment 2; for reviews, see Marques, Páez, et al., 2001; Marques & Páez, 1994). Marques, Abrams and Serôdio (2001, Experiment 3) showed that derogation of ingroup deviants as compared to similar outgroup members is reinforced when the status of the ingroup as a whole is insecure. Participants were categorized in artificial groups (“Pictorial” versus “Experiential”) on the basis of a bogus imagination test and asked to state their positions on an ethical continuum. The experimenter then informed participants that the purpose of the study was to establish the superiority of one type of imagination over the other. In a second session participants were informed either that the results in the first session were entirely clear and left no doubt about the ethical superiority of the participants ingroup (secure identity condition), or that those results were totally ambiguous as to which group was superior (insecure identity condition). Participants were then asked to judge normative ingroup or outgroup members whose positions on the continuum were socially desirable and matched participants’ positions or members whose positions were socially undesirable (deviant). The purported aim of these judgments was either to confirm (secure identity condition) or to disambiguate (insecure identity condition) the results obtained in the first session. The results indicated that participants upgraded normative ingroup members and derogated ingroup deviants as compared to equivalent outgroup members, and this effect was stronger in the insecure identity condition than in the secure identity condition. This finding is consistent with the idea that derogatory judgments of ingroup deviants serve to sustain a positive social identity. In another study (Marques, Abrams, & Serôdio, 2001, Experiment 1), psychology and law students examined answers purportedly given by other psychology or law students to a previous survey about a relevant issue (student initiation). Most of these students either showed agreement (norm validation condition) or disagreement (norm undermining condition) with student initiation. Participants then indicated the most pronorm and least pro-norm position they would accept about student initiation along a pro-anti initiation continuum. When the norm was challenged, participants reported a higher threshold of rejection as given by the least pro-norm position they found acceptable than when the norm was validated by strong consensus. However, the most accepted position did not vary across conditions (see Figure 19.1). This result is crucial to understanding the impact of normative positions on judgments of deviants. In the norm-undermining condition, participants considered the least accepted position for ingroup targets to be closer to the norm than the least accepted position of outgroup targets (see Figure 19.1). They also evaluated the ingroup targets holding the least accepted position more negatively than their outgroup counterparts, whereas normative ingroup targets were upgraded relative to normative outgroup targets (see Figure 19.2). Finally, when stating how strongly they would attempt to persuade deviant targets to change their opinion in a forthcoming discussion, participants reported higher willingness to persuade deviant targets in the ingroup norm undermining condition than in all other conditions.
274
THE PSYCHOLOGY OF GROUP PERCEPTION
FIGURE 19.1. Most accepted and least accepted normative positions as a function of targets group membership and norm validation vs. undermining condition (Marques, Abrams, & Serôdio, 2001, Experiment 1)
FIGURE 19.2. Evaluations of targets who adopted the most accepted and least accepted positions in the ingroup and the outgroup as a function of norm validation vs. undermining condition (Marques, Abrams, & Serôdio, 2001, Experiment 1)
In another study (Marques, Abrams, & Serôdio, 2001, Experiment 2), following a bogus test, half of the participants received feedback that categorized them in two opposed types of imagination (group condition) whereas others were simply informed of their personal imagination characteristics (interpersonal condition). Participants were then asked to indicate their position on an ethical continuum. In a second session participants in the group condition received false feedback about the ethical positions of other ingroup or outgroup targets. In the interpersonal condition targets were simply presented as being interpersonally similar to or different from participants in terms of imagination characteristics. All but one of the targets adopted norm-validating positions on the continuum (normative targets). These normative individuals were either highly consensual (high uniformity condition) or somewhat dispersed (low uniformity condition). The remaining target adopted a norm-undermining position (deviant). Participants then evaluated one normative target and the deviant and expressed how willing they were to influence these individuals to change their positions on the continuum. Consistent with the black sheep effect, participants evaluated the normative target more positively and the deviant target more negatively when they were presented as ingroup members than when they were presented as outgroup members or simply as individuals. More importantly, this result emerged most clearly when the ingroup lacked normative uniformity (see Figure 19.3). As was the
THE MAINTENANCE OF ENTITATIVITY
275
FIGURE 19.3. Evaluations of normative and deviant targets in interpersonal and group settings as a function of group uniformity (Marques, Abrams, & Serôdio, 2001, Experiment 2)
case in the previous study, participants were also more willing to influence deviant individuals to change their opinion in the ingroup low uniformity condition than in all other conditions. Three aspects of the results from the above studies are particularly worthy of attention. First, deviance per se does not necessarily generate norm-reinforcing responses (in the form of derogatory judgments and motivation to reintegrate the deviants). Such responses emerge more strongly when the social validation of a relevant dimension is endangered by lack of perceived consensus among group members. It seems that, when the group membership of targets is well established, what individuals care about is the extent to which ingroup norms are prescriptively validated or undermined. Second, such responses emerge primarily in settings where deviants are deemed to be ingroup members rather than in interpersonal or in outgroup settings. Third, derogation of ingroup deviants is not necessarily equated with the exclusion of such deviants from the group. Rather than ostracizing the deviants, there seems to be a strong motivation to influence these deviants in order to persuade them to adhere to the prescriptive norm, especially when the ingroup lacks uniformity.
276
THE PSYCHOLOGY OF GROUP PERCEPTION
The Role of Denotative and Prescriptive Norms on Judgments of Groups and their Members A series of studies conducted by Marques, Abrams et al. (1998) support the idea that category (denotative norm) differentiation and prescriptive norm differentiation are distinct processes that operate together. In this series of experiments deviance was manipulated in a way that was not confounded with general likability or social desirability. In each experiment participants first completed a bogus decision-making test in which they were asked to read a fictional murder story and then rank-order six characters in terms of their responsibility for the murder. In a later session, participants were informed that, based on the pretest, they had one of two decision-making styles (“Type X” versus “Type Y”). They were then told about the way people with this decision-making style should be expected to rank the six characters in the story (this corresponded to participants’ own rank ordering), and that people with the alternative imagination style should tend to use the opposite rank ordering. These manipulations defined the descriptive or category differences (X vs. Y) and the prescriptive norms for each category (rank ordering of story characters). Participants were presented with the rank-order judgments made by five group members, either from the ingroup or from the outgroup. Four of these members displayed the rank order linked to their category membership, the fifth member displayed a rank order that was more like the opposed pattern. Note that in these experiments the ingroup pattern and the participant’s own rank-order pattern was identical. This ensured that the relative differences between self and target group members were constant across conditions, and it also enabled us to distinguish effects of group membership and deviance from those of interpersonal similarity, because the latter variable was constant. Participants evaluated the ingroup as a whole, the outgroup as a whole, and each of the five target members (from either the ingroup or the outgroup). The first experiment (Marques, Abrams et al., 1998) illustrated that both denotative and prescriptive differentiation occurred, consistent with the SGD model. Participants judged the ingroup as a whole more favorably than the outgroup (denotative norm differentiation). However, participants also upgraded members whose responses were closer to the ingroup norm and derogated members whose responses were opposed to this norm, irrespective of whether they were ingroup or outgroup members (prescriptive norm differentiation). Three further experiments investigated the processes that increased or decreased denotative and prescriptive norm differentiation. Marques, Abrams et al. (1998, Experiment 3) manipulated whether participants referred to the prescriptive norm as a comparison standard. In the low prescriptive salience condition, participants were simply provided with the responses of the five ingroup or outgroup targets to the ranking task. In the high prescriptive salience condition, participants were additionally informed that although it was not necessary for belonging to a decision-making style, in general, people who belong to the (ingroup or outgroup) style should make a particular ranking, whereas those who belong to the other style, should rank the characters in the opposed way. The salience of prescriptive norms did not affect expressions of global ingroup bias, which was equally strong and significant in both conditions. Similarly, when the prescriptive norm was not salient, participants judged individual ingroup members more favorably than outgroup members, regardless of whether the members were normative or deviant. However, when the prescriptive norm was made salient, the pattern of findings from Experiment 1 reappeared; participants upgraded ingroup and outgroup members whose responses were closer to the ingroup norm and derogated ingroup and outgroup members whose responses were opposed to this norm. Therefore, salient cat egory differences resulted in denotative differentiation in all conditions, but prescriptive norm differentiation only occurred when participants were attentive to the prescriptive standards for the group. In Experiment 2 (Marques, Abrams et al., 1998), prescriptive norms were salient in all conditions, but participants believed their responses would be observed either by other ingroup members (ingroup
THE MAINTENANCE OF ENTITATIVITY
277
FIGURE 19.4. Favorability towards ingroup and outgroup group members as a function of accountability (Marques, Abrams et al., 1998, Experiment 2)
accountability condition) or by outgroup members (outgroup accountability condition). Previous research indicates that accountability increases self-focused attention (e.g., Abrams, 1990; Abrams & Brown, 1989). We therefore expected that accountability to ingroup members should be especially likely to enhance concern with ingroup prescriptive norms. As expected, derogation of the ingroup deviant was significantly stronger in the ingroup accountability condition (see Figure 19.4). Experiment 4 by Marques, Abrams et al. (1998) explored the hypothesis that intragroup differentiation in terms of prescriptive norms should be related to the participants’ level of identification with their group. Identification was measured immediately following categorization and then again after participants had evaluated the normative and deviant ingroup or outgroup targets. Consistent with our assumption that social identity motivates evaluative preference for ingroup and outgroup members who endorse prescriptive ingroup norms, prior identification with the ingroup was associated with evaluative differentiation between normative and deviant group members. In turn, consistent with the idea that prescriptive norm differentiation bolsters social identity, increased evaluative differentiation was associated with significant increases in identification from time 1 to time 2. Taken together, this series of studies support the SGD model and, specifically, the parallel operation of denotative and prescriptive differentiation and their association with the maintenance of social identity. Pro-Norm and Anti-Norm Deviance The above experiments are representative of analogous situations in which deviant ingroup group members may be the object of criticism and attempted influence. However, fundamental to the SGD model is the assumption that group members are concerned with the implications or meaning, rather than simply the presence and magnitude, of deviance from prescriptive norms. If deviance potentially undermines ingroup norms, then indeed it is likely to attract hostile reactions. Conversely, if deviance potentially validates those norms, it is likely to attract positive evaluations. In other words, we believe that group members may be tolerant, or even approving, of deviants whose differences from other group members mean they can contribute positively to the subjective validity of the group norm. To investigate this possibility we extended our research paradigm to include two types of deviance. Anti-norm deviance differs from the norm of a group in a direction that undermines or rejects the group’s position, and it may reduce the group’s distinctiveness relative to opposing outgroups. Pro-norm deviance, by contrast, differs from the norm of a
278
THE PSYCHOLOGY OF GROUP PERCEPTION
group in a direction that validates and supports its aims or ethos and may enhance the group’s distinctiveness relative to opposing outgroups (Abrams, Marques, Bown, & Henson, 2000). Whereas antinorm deviants may be considered by their group as traitors, pro-norm deviants may be viewed as idealists, loyal enthusiasts, or perhaps fanatics. To test these ideas, Abrams et al. (2000, Experiment 1) asked participants to evaluate a series of targets from their own gender group who were ostensibly being considered for promotion. The social characteristics of each candidate were depicted using a graphical display, such that the different candidates were all very similar in levels of competence, intelligence, politeness, and other features, but one candidate was more feminine, and another was more masculine than the other (normative) members. The magnitude of deviation from the normative members was objectively equivalent in each direction, and these differences were perceived accurately by participants. Participants regarded themselves as significantly more similar to the normative target than to either of the other targets, but they also felt the pro-norm target had more in common with the group and themselves than did the anti-norm target. We predicted that evaluations of these targets would reflect how much each provided support for ingroup gender norms vis-àvis opposing gender norms. Consistent with this idea, participants rated the normative members as more attractive than the pro-norm deviant, but rated the pro-norm deviant as more attractive than the anti-norm deviant. Moreover, these effects remained significant once ratings of perceived similarity between self and target were partialed out. Thus, although pro-norm deviants were disliked, they were tolerated more than anti-norm deviants, because pro-norm deviants were less undermining of ingroup norms. A further study (Abrams, Marques, Bown, & Dougill, 2002, Experiment 1), examined reactions to deviance in a commercial banking organization. Em ployees in a major U.K. offshore bank read descriptions of behavior by other ingroup workers. All participants read about a normative worker. Half the participants also read about an anti-norm deviant who was critical of the organization, refused to do overtime work, and so forth. The other participants read information about a pro-norm deviant who was obsessed with supporting the organization and would make special efforts to work additional hours, recruit new members, and so forth. As in Abrams et al. (2000, Experiment 1), evaluations of the anti-norm deviant were significantly more negative than those of the pro-norm deviant, even though they both were perceived as being equally different from the ingroup norm. Again, as in Abrams et al. (2000, Experiment 1), these effects remained significant even when perceived self-to-target similarity was used as a covariate. Moreover, more negative evaluations of deviants were significantly associated with prior identification with the organization. Taken together, Abrams et al. (2000, Experiment 1) and Abrams et al. (2002, Experiment 1) suggest that, when distinguishing among ingroup members, people are equally able to detect the magnitude of pro-norm and anti-norm deviance, but they reserve their most negative evaluative reactions for anti-norm deviants. Two further studies extended the above research to explicitly intergroup contexts and examined the idea that intergroup comparisons should make people more concerned both to ensure the validity of ingroup norms and to maintain ingroup entitativity (see also Yzerbyt et al., 2000). Abrams et al. (2000, Experiment 2) focused on attitudes about the number of asylum seekers that should be allowed entry to Britain each year. A pilot study established that psychology students believed the existing entry rate was about right. In the main experiment, psychology students read the results of national surveys that ostensibly had been conducted among psychology students or customs and immigration officers. They were told that psychology students wanted no change in the percentage of asylum seekers allowed to remain in Britain but that customs officers advocated a reduction in the numbers by 30%. Participants then viewed responses to several of the survey items by six respondents, either from the Psychology Survey or from the Customs Officers Survey. In fact, four target group members were normative in their opinions, the fifth was pro-
THE MAINTENANCE OF ENTITATIVITY
279
normative and the sixth was anti-normative. Across conditions and types of deviant the objective difference between normative and deviant targets was kept constant. Moreover, the ingroup and outgroup anti-norm targets actually expressed identical attitudes (i.e., that there should be a 15% reduction in the numbers of asylum seekers allowed to remain in Britain). As in the previous studies, the ingroup as a whole was evaluated more positively than the outgroup, and this difference was strongly associated with ingroup identification. Similarly, as in our previous studies, participants were accurate when asked to report the actual opinion position espoused by each target member. However, an interesting change emerged in the ratings of group member typicality. Specifically, pro-norm deviants were judged to be equally typical of their group as the four normative members. Only the anti-norm mem bers were viewed as being atypical. This suggests that judgments of typicality are made with reference to how much the target helps to validate prescriptive norms and not with reference to statistical typicality. In the intergroup context of the study it seems reasonable to suppose that typicality judgments reflected prototypicality as defined by the meta-contrast ratio in SCT (e.g., Haslam, Oakes, McGarty, Turner, & Onorato, 1995). Evaluations of the targets did not correspond to the specific attitudinal position they expressed. Ingroup normative members and ingroup pro-norm deviants were evaluated more positively than ingroup anti-norm deviants. The reverse pattern was obtained for outgroup targets (see Figure 19.5). In fact, the relatively ingroup-validating (outgroup anti-norm) deviant was evaluated more positively than the relatively ingroupundermining (ingroup anti-norm deviant) even though both targets expressed identical attitudes. Moreover, the more that participants identified with the ingroup, the more strongly they favored deviants who validated, as compared with deviants who undermined, the ingroup norm, regardless of whether the deviants were themselves from the ingroup or the outgroup. These results are highly consistent with the SGD model. Abrams et al. (2002, Experiment 2) conducted an analogous study in the context of the University of Kent’s policy for admission of students from outside Europe (“overseas students”). British universities charge a higher level of tuition fees to students from outside Europe, but accordingly they try to provide some advantages for these students, including privileged access to accommodation on campus and related schemes. Pilot studies confirmed that overseas students were happy with the situation but believed that further privileges would be justified. British students were also broadly acquiescent but believed that a reduction in privileges for overseas students would be appropriate. That is, although both groups were not discontented with the situation, they held diverging prescriptive norms regarding policies that would affect
280
THE PSYCHOLOGY OF GROUP PERCEPTION
FIGURE 19.5. Favorability to pro-norm, normative, and anti-norm group members as a function of group membership (Abrams, Marques et al., 2000, Experiment 2)
future generations of students. Participants were then presented with statements, ostensibly taken from a pilot study, made by three targets from each group. The statements concerned university policy for future cohorts of overseas students. From each group one target expressed the normative opinion of the group, another expressed an anti-norm position, and the last expressed a pro-norm position. For our manipulations of deviance, the anti-norm ingroup target and the pro-norm outgroup target expressed identical attitudes (more privileges for future members of the outgroup), that were equally divergent from the current norm (maintain the status quo). Conversely, the pro-norm ingroup and the antinorm outgroup targets also expressed identical attitudes (fewer privileges for future members of the outgroup). Consistent with Abrams et al. (2000, Experiment 2), anti-norm deviants were seen as significantly more atypical than normative members and pro-norm members. Typicality ratings of normative and pro-norm members did not differ. Thus, despite potentially strong demand characteristics to distinguish among all six targets on a single continuum, perceived typicality followed a principle of relative normativeness or prototypicality, independent of the actual attitude position expressed and independent of objective similarities among targets. This is consistent with the idea that typicality judgments are attributable to prototypicality defined in the intergroup context and not to absolute differences among targets or to whether the positions adopted by the target are ingroup validating per se. The pattern of evaluations was also consistent with that found in Abrams et al. (2000, Experiment 2). The pro-norm ingroup deviant was evaluated more positively than the normative ingroup member, and both were evaluated more positively than the anti-norm ingroup deviant. The reverse pattern was obtained for outgroup targets. Evaluations of the pro-norm ingroup and anti-norm outgroup deviant were equally positive. Moreover, we found that evaluations of ingroup validating (ingroup pro-norm and outgroup antinorm) deviants, relative to ingroup-undermining (ingroup anti-norm and outgroup pro-norm) deviants, were strongly associated with the extent to which participants rated the two types of deviants as differing in typicality. The more that participants focused on discrepancies between the prototypicality of deviant and normative members, the more they showed favoritism to (ingroup and outgroup) members who validated the ingroup norm. Across the Abrams et al. (2000, 2002) studies the evidence converges to show that, as group membership becomes more salient (i.e., as the context becomes more explicitly intergroup), people may engage in more intragroup differentiation in terms of prescriptive norms. For ingroup targets, anti-norm deviants are evaluated very negatively, but pro-norm deviants are often evaluated as if they are normative (in an intragroup setting) or even more favorably than normative members (in an intergroup setting). Even when
THE MAINTENANCE OF ENTITATIVITY
281
perceivers are well aware of the objective magnitude of deviance, their evaluations reflect the implications of deviance for the validity of ingroup norms and social identity. Additional evidence shows that evaluative differentiation within the ingroup seems to be associated with higher group identification (Branscombe et al., 1993; Castano, Paladino, Coull & Yzerbyt, 1999; Hutchison, 2000; Hutchison & Abrams, in press; Marques et al., 1988). However, it is important to note that evidence suggests this may be accompanied by a motivation to reintegrate deviants by gaining control over them rather than a motivation to reject them from the group (Marques, Abrams, & Serôdio, 2001). ENTITATIVITY AND SUBJECTIVE GROUP DYNAMICS As discussed above, previous theory and research (e.g., Lickel et al., 2000; Oakes et al., 1994; Sherman et al., 1999) seem to suggest that the presence of ingroup deviants should reduce the fit of existing categorizations, and also group entitativity, particularly if the deviants are anti-normative. However, our own research suggests that intragroup differentiation often goes hand-in-hand with intergroup differentiation and that both may function together as a means of sustaining the relative validity of ingroup norms (see also Castano et al., 1999; Yzerbyt et al., 2000). For example, Coull, Yzerbyt, Castano, Paladino, and Leemans (2001) found that group members who identify more strongly with their group are likely to devote more cognitive resources to processing information about deviants but nonetheless strongly sustain the group stereotype. Concordant with this, using entitativity measures from Lickel et al. (2000), Hutchison (2000) found that the presence of an ingroup deviant in a sports context actually increased the perceived entitativity of the ingroup relative to an outgroup, and this effect was particularly strong when the deviant was anti-normative and among group members who identified more highly. This suggests that the presence of deviants may elevate perceived ingroup entitativity, not because the deviants are excluded from the group, but rather because their salience also heightens the salience of ingroup norms (Marques, Abrams, & Serôdio, 2001). What might be responsible for this effect? We suggest that intragroup differentiation may contribute to, and may indeed provide psychological validity for, intergroup differentiation. Our major point is that the two types of differentiation are based on distinct criteria. Evaluations of the category as a whole seem to be largely based on denotative differences and the higher value attached to ingroups than to outgroups. By contrast, evaluations of individual group members seem to be based largely on prescriptive norm differences and the definition of social reality. Ultimately, the higher value of the group depends on validation of its definition of reality, and therefore these two types of differentiation must work in conjunction rather than in opposition to one another. CONCLUSIONS We believe that the SGD model can augment theory and research across a range of central topics in social psychology. As well as its obvious significance for in tra- and intergroup relations, prejudice, and social influence it has relevance to research on stereotyping and stereotype change (Hutchison & Abrams, in press; Hutchison, Abrams, & Viki, 2002), self and identity (Baumeister, 1999; E.R. Smith, Coats, & Murphy, 2001), attraction (Hogg, 1993), ostracism (Williams et al., 2000), self-regulation (Carver & Scheier, 1998), individual-group discontinuity phenomena (Insko et al., 1983), motivated cognition (Kruglanski & Webster, 1996), and small group process (Abrams, 1999b). In the SGD model, judgments of group properties, such as entitativity, homogeneity, and variability can be viewed not merely as perceptual outcomes of various objective differences within and between groups
282
THE PSYCHOLOGY OF GROUP PERCEPTION
but as qualities that are psychologically engineered and indeed managed. The SGD model embraces both intergroup and intragroup levels of representation as part of a dynamic process by which people sustain social reality for the self. It starts with the premise that people actively make sense of their social environment and seek a location within it. Therefore, self-regulation and regulation of the social environment are necessarily part of the same process. The model extends the Social Self-Regulation model to apply to judgments of ingroups and outgroups and their members and proposes that individuals may use different means (e.g., positive evaluations of certain outgroup members and negative evaluations of certain ingroup members) to reach the same ends (validation of ingroup norms). The model suggests that phenomena such as the black sheep effect are explainable in terms of an overarching goal to validate ingroup norms and ingroup reality. In the present chapter, we presented evidence that when people identify highly with a group, they do not show unqualified increased support for the ingroup over the outgroup. Their higher levels of global ingroup bias are accompanied by greater favorability toward ingroup and outgroup members who increase the relative validity of ingroup norms. To conclude, the subjective group dynamics approach offers several novel hypotheses that, in many respects, contrast with other established theoretical perspectives. Our superordinate aim is to promote the interpenetration and integration of theories that include the social-cognitive, intragroup, and intergroup levels (Abrams & Hogg, 1999; Operario & Fiske, 1999). We hope that the Subjective Group Dynamics model can make a positive contribution in that direction and will stimulate further research. AUTHOR NOTE The authors acknowledge grants from the Economic and Social Research Council (R42200034207 and R000223087) and the Portuguese Foundation for Science and Technology (Project PRAXIS XXI/P/PSI/ 13192/98), which supported some of the research reported in this chapter.
20 On the Advantages of Reifying the Ingroup EMANUELE CASTANO Graduate Faculty, New School University
INTRODUCTION As so eloquently argued by Hamlet, existence is usually conceived as an either-or question; things either exist or they do not. When it comes to social groups, however, the situation is not so clear-cut. Existence can be quantified. Groups can exist more or less. This is the view expressed some 50 years ago by social psychologist Donald Campbell. Building on principles of Gestalt psychology, Campbell (1958) attempted to define the conditions under which an aggregate of individuals begins to be perceived as a group. The term entitativity was coined, which referred to “the degree of having the nature of an entity, of having real existence” (p. 17). After remaining dormant for several decades, the concept of entitativity was recently resuscitated by social psychologists.1 In two influential reviews, Brewer and Harasty (1996) and Hamilton and Sherman (1996) convincingly used the concept of entitativity as a theoretical umbrella for an understanding of a series of phenomena in stereotyping and impression formation. Most of the research reviewed by these authors focused on the outgroup, or on groups of which the perceiver was not a member. Is the concept of entitativity useful when the object of investigation is the ingroup? In this chapter I provide some evidence that this is indeed the case. I review a series of findings that have emerged from a research program I conducted with colleagues in Europe and the United States. This evidence is organized into two main parts. In the first part, I focus on the idea that ingroup entitativity has an impact on ingroup identification.2 Findings from research studies involving broad political categories and smaller, real and artificial groups are reported. In the second part I offer a perspective according to which entitativity is a valued characteristic of the ingroup because it reifies it. Specifically, I suggest that it has a terrestrial and a celestial value: terrestrial, inasmuch as ingroup entitativity enhances the perception of the
284
THE PSYCHOLOGY OF GROUP PERCEPTION
ingroup as an agent that can guarantee security to its members; celestial, because social identification with entitative ingroups may help individuals projecting themselves in space and time, beyond the confines of their necessarily finite personal existence. INGROUP ENTITATIVITY AND INGROUP IDENTIFICATION In this first part I present a summary of the findings emerging from a series of studies conducted on a variety of groups in both Europe and the United States. These studies have in common a focus on the relation between ingroup entitativity and ingroup identification. They are organized into two sections. In the first section I present results of a research paradigm designed to test the impact of ingroup entitativity on ingroup identification. In these studies the ingroup of reference was the European Union (EU). In the second part, I summarize the results of a set of experiments carried out in the United States that focused on a particular antecedent of entitativity, common fate. In addition to investigating further the impact of entitativity on identification, these experiments also provided preliminary evidence for an empirical distinction between entitativity and identification measures. Manipulating Entitativity of Political Entities The research presented in this section was carried out using the EU as the group of reference. Our participants were Belgian students (i.e., European citizens) (Castano, Yzerbyt, & Bourguignon, in press). Our hypothesis was straightforward: increasing the perceived entitativity of the European Union would make European citizens more identified with it, while decreasing the entitativity of the European Union would trigger lower levels of identification. The rationale was that entitative ingroups are more attractive to individuals because they are perceived as entities having a real existence and thus having a greater capacity for action (Abelson, Dasgupta, Park, & Banaji, 1998; Castano, 1999; Yzerbyt, Castano, Leyens, & Paladino, 2000) and intentionality, as well as providing its members with a sense of security.3 In the next part of this chapter, which deals with the terrestrial value of ingroup entitativity, this rationale will be discussed in greater detail. Because of the specific context in which the experiments were conducted, an additional hypothesis referring to the a priori attitudes held by participants toward the European Union, was formulated. Specifically, we expected that the effect of the entitativity manipulation would be moderated by the initial attitudes held toward the EU. The European Union is a newly formed supraordinate entity, toward which European citizens hold very disparate attitudes ranging from enthusiasm to strong skepticism and aversion. Given that our dependent variable was the level of identification with the EU, it seemed reasonable to expect that individuals falling at the poles of the attitude scale would respond quite differently. Furthermore, given that our manipulation of entitativity would make the EU more or less appealing (for the reasons stated above), we also expected that individuals scoring at the poles of the attitude scales would react differently to the manipulation. This second point builds on Sherif and Hovland’s (1961) theory regarding attitude change, which suggests that the range of latitude for acceptance is quite small for those individuals holding polarized attitudes. Two processes may thus interfere with our manipulation. On the one hand, individuals holding extreme attitudes may accept the message conveyed in the manipulation but not change their level of identification with the EU because forces much stronger than the experimental manipulation drive their level of identification. Alternatively, individuals holding extreme attitudes may simply not accept the message conveyed in the manipulation. In other words, the perception of entitativity of the EU might not change
ON THE ADVANTAGES OF REIFYING THE INGROUP
285
because it is already extreme. Indeed, Euro-skeptics tend to perceive the EU as a meaningless invention of a group of bureaucrats—that is, as a nonexisting entity—whereas Euro-enthusi asts perceive it in quite opposite terms (Castano & Tousignant, 1999). In other words, a positive linear relation is likely to exist between attitudes towards the EU and the perception of EU entitativity.4 Accordingly, a measure of attitude toward the EU was included in all the experiments. The measure consisted of a list of traits from which participants selected those that best described their feelings when thinking of the EU as well as a measure of their overall attitude toward the EU. A composite attitude score was computed, and participants were categorized into three groups (negative, moderate, positive). For the reasons outlined above, we expected our manipulation to affect mainly, if not exclusively, those individuals holding moderate attitudes toward the EU. Building on Campbell’s (1958) conjectures and more recent theorizing on the mechanisms governing the representation of social groups (Turner, Hogg, Oakes, Reicher, & Wetherell, 1987), we selected four different manipulations of entitativity: common fate, similarity, salience, and boundedness. Each of these was used in a separate experiment. Our dependent variable was the level of identification with the EU as measured by a six-item identification scale (e.g., “I identify with other European citizens”, “It is important for me to be a citizen of the European Union,” “Being a citizen of the European Union is not part of my identity”). Of the four experiments, two included pre- and post-manipulation measures of identification (the identification scale was split into two subscales), while the remaining two included only post measures. In the first experiment (Castano et al., in press, Study 1) we focused on common fate. After the attitudes and pre-manipulation identification measures were taken, the manipulation of entitativity of the EU was introduced. Participants were presented with a video, produced by the European Commission and distributed among European citizens especially for educational purposes. The video consisted of a presentation of European history and stressed the fact that European countries shared a common fate within the European Union. As expected, the video triggered an increase in the level of identification, but this was the case only among participants holding moderate and, to a lesser extent, positive attitudes. No effect was observed among individuals holding negative attitudes. In the second experiment (Castano et al., in press, Study 2) we focused on perceived similarity among the European countries making up the European Union. In this experiment we focused on both the effect of increasing the perceived entitativity, and the effect of decreasing it, on the level of identification. Participants in two conditions received the same information about the EU, consisting of a table representing various characteristics of the 15 countries in the European Union (type of political system, number of chambers of parliament, etc.). The difference in the two conditions was in the instructions for looking at this information. In the high entitativity condition, participants were asked to identify similarities among the different countries, whereas in the low entitativity condition they were asked to identify differences. All participants were then requested to write a paragraph describing the EU. The manipulation produced an effect on the level of identification only among those individuals holding moderate attitude toward the EU. In the low entitativity condition these individuals showed a decrease in level of identification with the EU (M=−0.33), whereas in the high entitativity condition they tended to increase their level of identification (M=0.21), though the latter tendency was not significantly different from zero. In this second experiment we hypothesized that manipulating similarity among European countries would lead to the perception of the EU as more or less entitative, and that this, in turn, would lead to higher or lower identification with the EU. To test this hypothesis directly, we asked two independent judges to evaluate the paragraph written by participants. After introducing them to the notion of entitativity, we asked them to score each paragraph on a seven-point scale (1=not at all entitative; 7=very entitative). This score was
286
THE PSYCHOLOGY OF GROUP PERCEPTION
then used to compute mediational analysis on the data of the moderate-attitude group, the only one in which the manipulation had an effect on the identification.5 The results provided strong support for our hypothesis showing that the representation of the EU did indeed mediate the impact of the manipulation on the level of identification. The remaining two experiments carried out in this research program relied on a design with postmanipulation identification-measures only. In the first experiment (Castano et al., in press, Study 3) we focused on the salience of the ingroup. Building on self-categorization theory (Turner, 1975; Turner et al., 1987), we predicted that making the EU more salient would trigger higher levels of identification. One way to make the ingroup more salient comes from Oakes’s (1987) meta-contrast. The meta-contrast principle suggests that in an intergroup context, the ingroup will be more salient due to the presence of an outgroup. In contrast, in an intragroup context, the ingroup is likely to be nonsalient, due to its low relevance as a category for comprehending the situation. We therefore reasoned that in an intergroup context the EU would be perceived as more entitative and thus predicted higher levels of identification for the moderate-attitude group in this context. The results of an experiment in which the levels of identification with the European Union were measured supported our hypotheses. We compared the results of a questionnaire with no reference to non-European countries to one in which a non-European country was referenced (Canada). As anticipated, in the intergroup condition, moderate-attitude individuals displayed higher levels of identification than in the intragroup condition. In the final experiment we assessed the impact of the level of boundedness of the group (Castano et al., in press, Study 4). The manipulation consisted of depicting the EU as a well-bounded (high-entitativity) versus ill-bounded (low-entitativity) entity. Participants read a fabricated press article about the enlargement of the EU. The article included a brief section saying that the eastern EU borders were/were not clearly defined as well as a map in which these borders were/were not clearly defined. After reading the article, participants in both conditions were required to write a paragraph describing the EU. As in Study 2, two independent judges rated the extent to which the descriptions depicted the EU as entitative. The results confirmed our hypothesis and showed an effect of the manipulation on individuals holding moderate attitudes toward the EU. Of these participants, those who were confronted with a ill-bounded EU identified less strongly with the EU than those who were confronted with a well-bounded EU. Comparisons with the identification levels observed in the other attitude groups showed that the effect observed among moderate attitude individuals was due to a de-identification in the ill-bounded condition rather than an increase in the level of identification in the well-bounded condition. As in Study 2 of Castano et al. (in press), we used the judgments provided by the two judges as a mediator of the effect of the manipulation on the level of identification with the EU.6 The results of these analyses confirmed the mediational hypothesis and showed a significant reduction in the effect of the manipulation on identification, which, as a result, was not significant. Taken together, the results of the above experiments yielded clear support for the notion that increasing and decreasing the entitativity of the ingroup leads to an increase and decrease in the level of identification with the ingroup, respectively. Manipulating Entitativity in Artificial and Minimal Groups Ingroup entitativity and its links with ingroup identification were further explored in another set of experiments. One experiment was conducted using an artificially created group, and two others focused on real-minimal groups, a class of freshmen. These experiments aimed at replicating and extending the above findings, using different manipulations of common fate. They also made use of a translated and adapted
ON THE ADVANTAGES OF REIFYING THE INGROUP
287
version of an entitativity scale originally created for broad social categories (Castano, Yzerbyt, & Bourguignon, 1999). The entitativity scales used in both the artificial group experiment and the freshmen experiments comprised seven items designed to measure factors such as similarity between, and cooperation among, group members. Taking into account the specificity of the context, two items in the artificial group focused on distinctiveness, whereas in the freshmen experiments one item questioned whether the participants could think of themselves as a group.7 Identification was measured with a four-item or a sixitem identification scale (for the artificial group and freshmen experiments, respectively) derived from the scale used in Castano et al. (in press). An additional aim of these experiments was to test the discriminant validity of entitativity and identification. The Artificial Group Experiment. The manipulation of the common fate used in Study 1 of Castano et al. (in press) had the advantage of using realworld material, likely to have an impact on future generations of Europeans. However, the downside of this manipulation was the lack of control over the content of the video used for the manipulation. In conducting a further experiment using the minimal group paradigm we therefore decided to use a more neutral manipulation of common fate. In this experiment (Castano & Brewer, 2000) participants were told that they had been divided into two groups using the classic overestimator versus underestimator paradigm—actually everybody was classified as an overestimator. The manipulation of common fate was then introduced. The experimenter explained that participants were requested to sign up for another experiment. He asked them which group they belonged to and handed them a timetable on which they could sign up for a preferred time. In the common fate condition, about half of the slots were reserved for overestimators (i.e., the participants’ ingroup). In the control condition, the same slots were left unreserved (the same number of slots were available in the two conditions). After signing up, they were given a booklet that comprised the entitativity and the identification scales. The results were in line with the hypotheses, showing an effect of condition on the entitativity and identification measures. Compared to the control condition, participants in the common fate condition perceived the ingroup as more entitative and identified more strongly with it. A factor analysis with an oblique rotation revealed that the items measuring entitativity and identification loaded on two different factors, and that the factors were positively corre lated. Mediational analyses were also conducted to better assess the process through which our manipulation affected entitativity and identification. Two models, one in which entitativity served as mediator (and identification was the DV) and one in which identification served as mediator (and entitativity was the DV), found support and thus prevented drawing strong conclusions about the precise causal relation that might hold between the two variables. The Minimal-but-Real Group Experiments. The two manipulations of common fate discussed above focused on slightly positive or neutral common fate. What happens when the common fate shared by group members is a negative one? We investigated this hypothesis in two studies conducted with freshmen at Ohio State University (OSU) (Castano, Brewer, & MacDonald, 2000). Central to the manipulation in both experiments was the senior comprehensive exam, an exam that several universities are considering introducing at the end of the undergraduate program. One can assume that the vast majority of undergraduates would regard this as a negative event. Participants in both experiments were introduced to the concept of the senior comprehensive exam, then were told either that their class would be the first to take the exam (negative common fate condition) or that a decision had yet to be taken as to the implementation of the exam at OSU (control condition). Since results of the two experiments were virtually identical, we present the findings for the combined data set (cf. Castano et al., 2000). The manipulation affected both the perception of ingroup entitativity and the identification with the ingroup with negative common fate triggering higher entitativity and higher identification (see Figure 20.1). Mediational analyses
288
THE PSYCHOLOGY OF GROUP PERCEPTION
FIGURE 20.1. The effect of negative common fate on entitativity and identification (From Castano, Brewer, & MacDonald, 2000)
yielded support for both models—that is, the two models in which entitativity and identification are switched in their role of me diator and DV. We then compared the fit of these two models through a structural equation modeling procedure. The model in which entitativity mediated the impact of the manipulation on identification better fit the data than the model with identification as a mediator (and entitativity as the dependent variable). Finally, we computed confirmatory factor analyses on this combined data set. The picture that emerged from these analyses confirmed that the two-factor solution (with entitativity and identification items loading on two separate factors) better fit the data than the single-factor one (with both entitativity and identification items loading on the same factor). This provides further support for an empirical, as well as conceptual, distinction between entitativity and identification. SUMMARY The first part of this chapter focused on a series of experiments exploring the impact of ingroup entitativity on the level of identification with the ingroup. We presented data from experiments conducted with large and meaningful political identities (Castano et al., in press), with artificially created groups (Castano & Brewer, 2000), and with minimal-but-real groups (Castano, Brewer, & MacDonald, 2000). These experiments also varied in the way entitativity was manipulated. The findings in all the experiments, however, were remarkably uniform; all showed the expected impact on identification. More specifically, similarity among ingroup members, salience of the ingroup, ingroup boundedness, and positive, neutral, and negative common fate all proved to influence the level of identification with the ingroup. Where possible, mediational analyses were conducted to better understand the relation between perceived entitativity and identification with the ingroup. These analyses yielded mixed results. In the two experiments in which perceived entitativity was obtained by two independent judges’ ratings of a paragraph written by participants, clear support was found for our conjecture that increasing the perceived entitativity of the ingroup leads to an increase in the level of identification with the ingroup. In other cases in which perceived entitativity was obtained directly from participant’s answer to the entitativity scale, however, evidence for the mediating role of perceived entitativity was less compelling.
ON THE ADVANTAGES OF REIFYING THE INGROUP
289
The studies in which clear support for the mediational effect of entitativity emerged differed from the others in yet another important way. Whereas in the former the entitativity of the ingroup was directly manipulated by manipulating the way the ingroup was represented (Castano et al., 1998, Studies 2 and 4), in the latter entitativity was manipulated by varying the degree of common fate among group members (Castano & Brewer, 2000; Castano, Brewer, & MacDonald, 2000). Although similarity, clarity of boundaries, and common fate all appear to impact on entitativity, the effect of common fate is more “distal” than either similarity or clarity of boundaries. As a consequence, when manipulating common fate, it might be more difficult to observe the mediational effect of entitativity on identification because the effect of the manipulation on the two variables might occur simultaneously. The manipulation of the similarity and boundaries of the ingroup immediately influences the representation of the ingroup; it is only subsequently that this affects level of identification. Receiving information about common fate might entail a slightly different process, causing an immediate increase of entitativity and identification. It also seems reasonable to argue that the direct manipulation of the representation of the group may be considered “cold” cognition in contrast to the “hot” manipulation of common fate, especially in the experiments on the impact of a negative common fate. At this stage, however, these are speculations that need to be corroborated by further empirical research. Where appropriate, we tested the discriminant validity of the concepts of entitativity and of identification. Although the two concepts were always significantly correlated, the items used to measure them loaded on different factors. Moreover, confirmatory factor analyses yielded evidence for the superiority of a twofactor solution over the one-factor solution. It thus seems that entitativity and identification are related but distinct concepts. THE VALUE OF INGROUP ENTITATIVITY In this section I examine the advantages associated with the process of reification of the ingroup. The nature of these advantages provided the criterion used to organize the discussion into two parts: the terrestrial and the celestial value of ingroup entitativity. The Terrestrial Value of Ingroup Entitativity In the perspective on ingroup entitativity offered in this chapter, I argue that increasing the entitativity of the group leads to its reification. As a result of this process of reification, the differences between the individual-unit and the group-unit are reduced. Specifically, the reified group, like the individual, comes to possess intentionality, defined as the presence of its own goals and agenda. The terrestrial value of ingroup entitativity derives, among other things, from the fact that entitativity brings about a perception of the ingroup as having intentionality.8 Also, once reified, groups are likely to be judged as more secure. From an evolutionary perspective it seems clear that group life provided individuals with a greater sense of security deriving from the protection against rival groups (Wrangham, 1979), the reduction of predator pressures (Alexander, 1974), and the improvement of defense strategies (Van Schaik, 1983; see also Sedikides & Skowronski, 1997). Brewer and Caporael (1990) argued that the social group could constitute the selection environment for human evolution at the individual level. If this is indeed the case, to be evolutionarily functional the group needs to foster internal organization and coordination efforts, communication, role differentiation and boundary definition.9 In other words, it needs to be entitative. The sense of security that a group can provide to its members may help them in dealing with an immediate physical threat as well as a more remote one. An ingroup provides its members with a sense of
290
THE PSYCHOLOGY OF GROUP PERCEPTION
security because it encompasses and protects them. Each members perception of this condition is likely to be accentuated when in his or her mind the ingroup is reified. In the context of international relations (IR), for instance, an increased sense of security may derive from the reification of one’s own nation. This rationale led Simona Sacchi and me to hypothesize that the perception of intentionality and security would be greater among highly entitative than minimally entitative groups. Some preliminary empirical evidence in support of this hypothesis comes from two studies, one correlational and one experimental (Sacchi & Castano, 2002). Both studies were conducted in an IR framework and focused on the ingroup. We began by collecting correlational data among undergraduates at Ohio State University (Sacchi & Castano, 2002, Study 1). Participants (American citizens) were provided with a questionnaire in which several variables were measured. First, they responded to a shortened version of the entitativity scale for the United States (i.e., the ingroup) developed by Castano, Yzerbyt, and Bourguignon (1999), adapted to fit the specific context of this experiment. (The items referred to similarity, cooperation, common fate, and common past.) Then identification was measured via a four-item identification scale, adapted from previous studies (e.g., I identify with Americans). Participants were further requested to evaluate the U.S. on several dimensions, including intentionality (The U.S. is following a well-thought-out master plan; The U.S. takes control of the situation and acts following its own intentions) and security (e.g., The U.S. is proving to be stable and strong; It is difficult to jeopardize the commercial supremacy of the U.S.). The variables showed an interesting set of correlations (see Table 20.1). The more participants perceived the United States as entitative, the more they perceived intentionality behind its actions. Notably, entitativity and intention TABLE 20.1. Intercorrelations between Entitativity, Identification, Intentionality, and Security Entitativity
Identification
Entitativity 1.00 .46*** Identification 1.00 Intentionality Security ***p<.001; **p<.01; *p<.05. (From Sacchi and Castano, 2002, Study 1)
Intentionality
Security
.35* .42** 1.00
.34* .24 .35** 1.00
ality also correlated with the perception of security in the international arena. The more entitative and the more characterized by intentionality the U.S. was perceived to be, the more secure their position was seen in the international arena. In a second study we manipulated the entitativity of the ingroup. For the manipulation of the ingroup entitativity in an international arena, that is of participant’s own country, we relied on a role-play methodology (Sacchi & Castano, 2002, Study 2). We also manipulated the relationship between the participant’s own country and another fictitious country, which as a result could be either cooperative or competitive. The participants were told the concept of the role-play, presented with a contrived international scenario, and asked to imagine that they were citizens of the imaginary country, Arland. In the first part of the experiment, although Arland (i.e., the ingroup) was illustrated on a map that depicted the land masses of other countries, no explicit reference was made to any other country. Only a description of Arland was included. The nature of the description varied depending on the particular condition. In the high entitativity condition, Arland was presented as a country with a long history, characterized by political and economic harmony, and with a high degree of similarity among its inhabitants. As well, a map of Arland was provided
ON THE ADVANTAGES OF REIFYING THE INGROUP
291
FIGURE 20.2. The effect of entitativity on intentionality, security, and threat (From Sacchi and Castano, 2002, Study 2)
in which there were no internal borders between regions. By contrast, in the low entitativity condition, Arland was described as having a diverse political, cultural, and economic history and population, and the map of Arland depicted the boundaries between six differently colored regions. Participants were then asked to respond to a series of questions assessing the perception of ingroup entitativity, identification, intentionality, and security, using the same items as in Study 1, which had been adapted to the role-play context (note that two negatively phrased items were also added to the scale of intentionality). After collecting these measures, another country (Yzet) was introduced to create an intergroup context. Yzet was described either as an enemy (competitive condition) or an ally (cooperative condition) of Arland and the perception of threat from this country was assessed (e.g., “The creation of a Yzet army might jeopardize the power of Arland in the international arena”). The results of this experiment yielded support for our hypotheses. Compared to the “citizens” of a marginally entitative Arland, the citizens of a highly entitative Arland identified more strongly with Arland, perceived greater intentionality, felt more secure in the international arena, and perceived less threat from the other country (see Figure 20.2). No interaction, but an additive effect, was observed between the manipulation of entitativity and the manipulation of the relations between the two countries on this latter dependent variable. The correlation between these variables replicated those observed in the study focusing on the United States. The Celestial Value of Ingroup Entitativity. If we accept the idea that part of our identity as human beings derives from our membership in social groups (Tajfel, 1978, 1981; Tajfel & Turner, 1979), we may well accept the idea that individuals participate in the existence of the group. However, the group exists beyond the individual, transcending the individual both spatially and temporally. By allowing the individual to go beyond his or her personal death, it provides a sense of transcendence (Anderson, 1991; McDougall, 1920/1973; Kelman, 1969; A.D.Smith, 1995). This is what I call the celestial value of ingroup entitativity. The more the ingroup is reified, the more it may serve as a vehicle for transcendence. It follows that individuals may increase ingroup entitativity and strengthen their links with the ingroup when they most need to transcend themselves as unique individuals, namely, when they are confronted with the idea of their own death.
292
THE PSYCHOLOGY OF GROUP PERCEPTION
Reminding individuals that they will die is the specialty of Jeff Greenberg, Tom Pyszczynski, and Sheldon Solomon. In the last decade, these scholars have carried out a considerable number of experiments in which participants were asked to think about their own physical death (for an early illustration, see Rosenblatt, Greenberg, Solomon, Pyszczynski, & Lyon, 1989). This manipulation, known as mortality salience, has been used in the framework of terror management theory (TMT; Greenberg, Pyszczynski, & Solomon 1986; Greenberg, Solomon, & Pyszczynski, 1997; Solomon, Greenberg, & Pyszczynski, 1991). TMT builds primarily upon the work of antrophologist Ernst Becker. Becker argued that the combination of our instinct to survive and our awareness of the inevitability of death gives rise to the potential for a paralyzing terror that would make life impossible (Becker, 1962, 1973, 1975). In order to cope with such anxiety, human beings need to have faith in a cultural system and to believe that they are living up to the systems values (Greenberg et al., 1986). Cultural worldviews and self-esteem—“the perception that one is a valuable member of a meaningful universe” (Greenberg et al., 1997)—are postulated to serve as the two mechanisms that buffer the anxiety arising from our awareness of the inevitability of death. An impressive amount of empirical evidence exists in support of this hypothesis (for a review see, Greenberg et al., 1997). Merging the insights from TMT, intergroup relations literature, and the emerging theoretical and empirical work on entitativity, my colleagues and I reasoned that identification with an entitative ingroup might provide individuals with yet another mechanism to prevent the anxiety arising from the awareness of personal death. We thus conducted an experiment utilizing the TMT mortality salience manipulation (Castano, Yzerbyt, Paladino, & Sacchi, 2002). Participants were randomly assigned to either the experimental condition (in which the idea of death was made salient by asking them to write a paragraph about their own death) or to a control condition (in which they were asked to write a paragraph about their thoughts when reading a book). A questionnaire including measures of ingroup entitativity, ingroup identification, and ingroup bias, allegedly unrelated to the previous exercise was then handed to participants. The experiment was carried out with Italian undergraduates at the University of Padua. The ingroup was Italians and the outgroup was Germans. If identifying with an entitative ingroup gives individuals a chance to escape, at least symbolically, from their own death, then we should observe greater ingroup entitativity and greater ingroup identification in the mortality salience condition. This was exactly the pattern that emerged from the results. Making a bridge between previous results by Harmon-Jones, Greenberg, Solomon, and Simon (1996), who showed the impact of making mortality salient on ingroup bias, and by L.Gaertner and Schopler (1998), who showed the impact of ingroup entitativity on ingroup bias, mediational analyses showed that the impact of mortality salience on ingroup bias was mediated by the perception of ingroup entitativity. Once again the data confirmed that entitativity and identification are related but distinct concepts. Indeed, a confirmatory factor analysis with a two-factor solution better fit the data than a single-factor solution (Castano, Yzerbyt, Paladino, & Sacchi, 2002). SUMMARY In this section, I focused on the value of ingroup entitativity. I proposed that group entitativity holds a terrestrial and a celestial value for group members. On the terrestrial side, building upon the idea that entitativity leads to the perception of the group as an entity having a real existence (Campbell, 1958), I argued that an entitative group is likely to be perceived as possessing greater intentionality and as providing its member with an enhanced sense of security. The link between entitativity and intentionality, which found empirical evidence in two studies (Sacchi & Castano, 2002), is in line with an argument advanced by Abelson et al. (1998). These authors have suggested that entitativity goes hand in hand with group activity, which “elicits
ON THE ADVANTAGES OF REIFYING THE INGROUP
293
an inference that the actor is engaged in active pursuit of goal satisfaction” (p. 248). An ingroup with such characteristics is likely to attract members more than a group that lacks them (cf. S.J.Sherman, Hamilton, & Lewis, 1999) . The idea that the reification of the ingroup goes hand in hand with the perception of security by its members also found support in our studies. Perception of security was indeed correlated with, and influenced by manipulation of, group entitativity. Further, in an IR, role-play setting, participants also felt less threatened by an outgroup when their ingroup was highly entitative, compared to when the ingroup was minimally entitative (Sacchi & Castano, 2002, Study 2). Both intentionality and a sense of security might be related to the idea that an entitative group is more effective. If in an intergroup context an effective group is likely to be perceived by its members as more secure than a less effective group, the members of the other group(s) are likely to perceive it as more threatening. Data from Dasgupta, Banaji, and Abelson (1999) are revealing in this regard. Participants in their experiment were presented with a cover story describing computer-generated humanoid creatures, called Gs, that were created for a new science fiction film (Dasgupta, Banaji, & Abelson, 1999; Experiment 2). The participants were advised that while the impression formation task would focus on the Gs exclusively, two kinds of creatures, the Gs and the Hs, would appear in the film. Results showed that when the Gs were presented as more entitative, they were expected to act more negatively toward Hs. Note, however, that these findings were qualified by the results of two experiments by Castano, Sacchi, and Gries (in press), who found that the actions of entitative enemy groups are perceived to be more harmful for the ingroup than actions from enemy groups that were low in entitativity. The same experiments also yielded evidence that actions of ally groups are judged to be less harmful when these are high rather than low in entitativity, leading Castano et al. (in press) to speculate that group entitativity might lead to polarized judgments. If we consider that an ally is almost an ingroup, these results are consistent with the findings described above showing the impact of ingroup entitativity on the sense of security. Further research is needed to address this point specifically. I would expect that more entitative ingroups would be perceived as more effective or more able to reach their goals (cf. Yzerbyt et al., 2000; for evidence of the reverse causality see Corneille, Yzerbyt, Rogier, & Buidin, 2001). Possessing goals, intentionality, and effectiveness, we have seen, characterizes a highly entitative group more than a less entitative group. These are also characteristics of individuals. Consistently, entitative groups, unlike nonentitative groups, are apprehended and judged in a similar way to that in which individuals are (McConnel et al., 1994; 1997; Welbourne, 1999) and the behavior of individuals belonging to highly entitative groups is more likely to be attributed to enduring group dispositions than that of members of less entitative groups (Yzerbyt, Rogier, and Fiske, 1998). In light of the terrestrial advantages of belonging to an entitative ingroup outlined above, it becomes clearer why group members displayed greater levels of identification when they were lead to believe that their ingroups were highly entitative. Although further research may well demonstrate empirically that some of these variables might play a mediating role in the impact of ingroup entitativity on ingroup identification, the preference for entitative ingroups is not necessarily due to a conscious and well-thought-out analysis of costs and benefits. As human beings, or even better as social animals, we know that entitative ingroups are better without necessarily knowing why (cf. Caporael, 1997). In addition to having a terrestrial value, I suggested that ingroup entitativity also possesses a celestial value. The social identity that membership in a social group provides to individuals allows them to symbolically project themselves through space and time. In this way, increasing the extent to which the ingroup is perceived as having a real existence (i.e., to reify the ingroup) is likely to release the individual from the sense of finitude of his or her own life (Castano, 1999; Castano, Yzerbyt, Paladino, in press;
294
THE PSYCHOLOGY OF GROUP PERCEPTION
Castano, Yzerbyt, Paladino, & Sacchi, 2002). A question that remains to be answered, however, is the exact mechanism through which group identification might serve as an anxiety-buffering mechanism. In the work presented above, it is proposed that identification with an entitative ingroup may buffer anxiety because it allows a shift from a personal, mortal identity to a social, immortal one. Quite a different rationale has recently been proposed by Dechesne and his colleagues (Dechesne, Janssen, & van Knippenberg, 2000; Dechesne, Greenberg, Arndt, & Schimel, 2000) in investigating the moderating role of ingroup status (or positivity) on the effect of making mortality salient on ingroup identification. Results suggest mortality salience increases identification with ingroups that can positively contribute to the individual’s self-esteem, but decreases identification for those ingroups that may reflect negatively on self-esteem. In the authors’ view, identification with the ingroup is considered functional to the enhancement (or protection) of self-esteem, i.e., the primary anxiety-buffer mechanism in TMT. This is not incompatible with the perspective we offered in our work. To better understand their interplay it would be fruitful to look at the dynamics of identification and deidentification within different types of groups (e.g., common identity vs. common-bond groups, Prentice, Miller, & Lightdale, 1994), which are alleged to serve different functions for the individual. Importantly, instead of (or in addition to) being a defense mechanism, ingroup identification may follow from a fundamental human need to belong or to be assimilated to others, as postulated by Baumeister and Leary (1995), Brewer’s (1991) optimal distinctiveness theory, and by Deci and Ryan’s (1991) selfdetermination theory (among others). These perspectives, similarly to TMT, provide a broad picture within which group behavior can be understood. Research that merges distal and proximal causes of group perception and behavior, as we have attempted by combining insights from terror management theory and social identity theory, will not only provide empirical tests of the applicability of general theories of human behavior, but also, by identifying fundamental needs inform us about the strategies for improving intergroup relations that are likely to be most successful. As in part one, part two also focused on the entitativity of the ingroup. Many of the insights that have served to elaborate the arguments presented here, however, come from research conducted on groups out there, or on outgroups. I have suggested that group entitativity is likely to produce a radically different set of outcomes depending on whether it is perceived in the ingroup or in the outgroup. To summarize, an entitative outgroup is threatening, whereas an entitative ingroup is reassuring. Research on ingroup entitativity and on outgroup entitativity is thus likely to be cross-fertilizing. Theoretical and empirical research assessing the interaction between ingroup and outgroup entitativity will certainly help better the understanding of this intruiguing aspect of group perception and intra- and intergroup behavior. INGROUP REIFICATION AT WORK: NATIONAL AND SUPRANATIONAL IDENTIFICATION IN TODAY’S EUROPE A context in which the process of ingroup reification appears to be at play with great power is the construction of national identities (Reicher & Hopkins, 2001). This is done, for example, by reference to a common past. Curiously, this reference need not be glorious. The “Je me souviens” which appears on the car registration in Quebec refers to the battle on the Plains of Abraham in which the Quebecois were defeated by the English. In these cases, it might be the promise of a future vindication that holds people together. Another example is the “Kosovo Polje,” the epic battle the Serbs lost to the Ottoman Turks in Kosovo, symbol of Serbian ethnic pride.10 As convincingly illustrated by Billig (1995), the process of community-building at the national level is a continuous one, so much embedded in every-day activities that it disappears in the background. References to the nation as the “motherland” or to the nation’s “founding fathers” are only some examples of the
ON THE ADVANTAGES OF REIFYING THE INGROUP
295
process of anthrophomorphisation that depicts nations as organisms with their own intentions, goals, and capacity for action. As we know, individuals are ready to die for the survival of this organism. However irrational this death may appear to some, it is, in the eye of the self-sacrificing individual, a “correct death” (P.L.Berger & Luckmann, 1967). The importance attached to national identity can be seen at work in many instances, some of which are characterized by extreme violence. Others are, happily, more peaceful, as in the case of Western Europe today. It is in the context of European integration that my colleagues and I have conducted interdisciplinary work on the specific issue of European identity (Castano, 2003; Castano & Tousignant, 1999; Castano, Yzerbyt, & Tousignant, 2000; Sindic, Castano, & Reicher, 2001). Since the end of World War II, Western European countries have been involved in a process of integration that has led to the creation of a supraordinate entity known successively as the Common Market, European Economic Community, European Community, and, more recently, the European Union. The modification of the name testifies to the changes, over the decades, in the nature of this entity. In the early days of integration the union was exclusively an economic one. Since the mid-eighties, however, the EU has become increasingly politicized. Concurrently, the process of integration originally guided by a neofunctionalist (Haas, 1958), elite-based logic, is striving for popular support. The dichotomy of economic versus identity concerns parallels the taxonomy proposed by Kelman (1969) that accounts for personal involvement in national systems. Kelman suggested that two sources of attachment (sentimental vs. instrumental) and three manners of integration into a national system (ideological vs. role-participant vs. normative) give rise to six different patterns of integration. Of particular interest here are the two patterns emerging from an ideological mode of integration. Ideological integration is likely to occur during the consolidation phase, that is, the phase in which the European Union now finds itself. Kelman argued that when the source of attachment in this phase is instrumental, the commitment of the individual is to institutions that promote the needs and interests of the population. This was the strategy adopted by European policy makers in the 1950s and ’60s. When the source of attachment is sentimental, however, the commitment is to cultural values reflective of national identity. Institutions, in this case, are supported not (only) because they provide satisfaction of material needs, but (also) because they are perceived to represent the community. The key point is the recognition of the homogeneous features within the community. However, I argued elsewhere that the critical variable might be the process of reification of the community rather than homogeneity per se (Castano, in press). The experimental research that my colleagues and I have conducted which focused on the European Union provides preliminary evidence in support of this conjecture (Castano et al., in press). Clearly, the idea of reification of political communities is often perceived as highly problematic: it raises the spectrum of social engineering. In today’s Europe this is also often debated because of the presence of strong national feelings, the basis of which may be undermined by the emergence of this supraordinate entity. It is my opinion that, inasmuch as political identities, like all social identities, are socially constructed, this is a false problem. As Reicher and Hopkins (2001) have said, to talk of an “authentic” national identity makes little sense. A fortiori, to define a possible European identity as “fake” makes no more sense. It follows that the process of reification of the European Union, and its potential long-term consequences on the self-perception of future generations of Europeans in terms of their political allegiances to “territorial identities” (Marks, 1999), is a matter of political struggle rather than of genuine versus fake, of what we “want,” rather than of what “is” (Castano, 2002). Whether or not, and to what extent, it is possible to create a sense of community without pretending that this is the only way things are or can be is an interesting and important question.
296
THE PSYCHOLOGY OF GROUP PERCEPTION
In addition to the issue of the legitimacy and the possibility of reifying the EU, and of its inward consequences (i.e., for European themselves), the question emerges as to the impact that such a process might have on the discrimination against non-European? If we look at the process of group reification as the ultimate form of category differentiation, we may be tempted to conclude that reifying the EU would lead to greater discrimination toward outgroups. The notion that increased category salience automatically leads to discrimination is traditionally, erroneously, attributed to Social Identity Theory. Reicher (in press), for instance, has recently reminded us not only that SIT has never made such a prediction, but also restated the conditions under which categorization is likely to lead to intergroup hostility and conflict (see also Turner, 1999; McGarty, 2001). Similarly, Jetten, Spears, and Manstead (1997) have shown that group norms can make pro-social behavior more rather then less likely to occur in an intergroup context. Comparing the effect of a prime of multiculturalism versus color-blind ideology on group perception, Wolsko, Park, Judd, and Wittenbrink (2000) come to the same kind of conclusions. Indeed, their findings suggest that increasing category salience, though causing stronger stereotyping, may even trigger more positive evaluations of the outgroup.11 All in all, though social psychology has long looked at group behavior as inherently problematic (and not without reason, cf. Insko & Schopler, 1998), the reification of an ingroup, although it has the potential to do so, does not necessarily trigger outgroup discrimination and intergroup conflict. Social groups, and the social identities that are linked to them, are not useless social constructions, but meaningful frameworks for understanding the world, and possibly an inherent characteristic of human life (Caporael, 1997; Castano, 1999; Yzerbyt et al., 2000). The question therefore is how to promote cooperation and mutual respect between social groups, not how to manage the improbable task of making the world “group-blind.” CONCLUDING REMARKS Does entitativity exist? The research that my colleagues and I have conducted on the concept of entitativity, its determinants and its consequences, suggests that entitativity, indeed, exists. It also indicates that, for a group, being entitative is to exist. In the present chapter I summarized the findings emerging from this research and presented a perspective within which entitativity is a key variable for the understanding of the process of group reification, its causes and its consequences. Consistent with the idea that this is a valued characteristic of the ingroup, increasing the entitativity of the ingroup was shown to impact on the level of identification with the ingroup. Also, under circumstances that are supposed to increase the need for social identification, the ingroup is perceived as more entitative. Although much more empirical research is needed to better our understanding of these intriguing and important phenomena, in the final section of this chapter I attempted to use these social psychological contexts for an analysis of the process of national and supranational identification in contemporary western Europe. In so doing, I realized, once again, that reality often goes well beyond fiction. Indeed, to forge common identities policy makers have made use of strategies very similar to those we used in our laboratory experiments. Analyzing these strategies in the context of a colonial identity construction for Indonesian inlanders, the historian and political scientist Benedict Anderson wrote: “Their [Inlanders’] common experience, and the amiably competitive comradeship of the classroom, gave the maps of the colony which they studied (always coloured differently from British Malaya or the American Philippines) a territorially specific imagined reality which was every day confirmed by the accent and physiognomies of their classmates.”12
ON THE ADVANTAGES OF REIFYING THE INGROUP
297
NOTES 1. See McGarty, Haslam, Hutchinson, & Grace, 1995, for the first empirical test of Campbell’s idea, as well as Lickel, Wieczorkowska, Lewis, Sherman, & Hules, 2000. 2. Identification is defined with Turner (1987) as “the centrality and evaluative importance of group membership in self-definition” (p. 55). 3. See Menon, Morris, Chiu, and Hong (1999) for an excellent discussion of cross-cultural differences in the perception of agency in groups. 4. It could also be argued that individuals holding extreme negative attitude see the EU as very entitative, but this is one more reason for them to dissociate themselves from the EU. Indeed, for them the EU might be considered an outgroup (cf. Castano, Sacchi, & Gries, in press; Dasgupta et al., 1999); 5. Only a main effect of the manipulation was observed on entitativity scores, suggesting that the negative and positive attitudes of individuals did not change their level of identification because forces stronger than the manipulation of entitativity were at work for them. 6. As for Study 2 of Castano et al. (in press), only a main effect of the manipulation was observed on entitativity scores. 7. In the context of an artificially created group this item might have created suspicion with respect to the overestimator/underestimator distinction. 8. If discrepancy exists between the goals of a member of the group and the goals of the group as a whole, any factor like entitativity that might increase the sense of group intentionality might come across, for this member, as negative rather than positive news. This is however a specific case. In this context I focus on the average situation in which the members of the ingroup are content with being in the ingroup and do not perceive any great discrepancy between their goals and the group goals. 9. The importance of role-differentiation and internal organization has been discussed, respectively, by S.A.Haslam (2001) who refers to the concept of organic pluralism in an organizational context, and by Hamilton, Sherman, and Lickel (1998). 10. Although these are examples of lost battles, the interpretation of the events and the process of meaning making that has followed is obviously more important than the actual outcome of the battle. In the case of Serbia, for instance, it is narrated that Serbian Tsar Lazar opted the heavenly instead of the earthly kingdom by “accepting” that his army should be defeated by the Turks (cf. Anzulovic, 1999). I am indebted to Steve Reicher for suggesting this reference. 11. I share most of the ideas put forward by Wolsko and his colleagues. However, I have some reservations about their findings, inasmuch as they may depend on the strong experimental demand that in my opinion is likely to be felt by participants in their experiments. The authors acknowledged but dismiss this possibility in the discussion. 12. B.Anderson (1991, p. 122). An obvious question is how successful this strategy has proved to be, given the process of devolution that characterizw Indonesia today.
AUTHOR NOTE This chapter is the product of several years of collaboration with friends and colleagues, Marilynn Brewer, Maria-Paola Paladino, Nathalie Tousignant, and Vincent Yzerbyt, and brilliant students, Simona Sacchi and David Bourguignon. Many thanks to Olivier Corneille, Nick Haslam, and Craig McGarty for their precious comments on an earlier draft of this chapter and to Mary Hoeveler and Tracy Solomon for editing suggestions.
21 Uncertainty and Extremism Identification with High Entitativity Groups under Conditions of Uncertainty MICHAEL A.HOGG University of Queensland
Read any newspaper. Surf the current affairs channels on television or the internet. What do you find? Along with the daily diet of sport and gossip, you will find reports of the activities of people who belong to what we might call “totalist” groups—since September 11, 2001, this has particularly been the case. These are people who appear to identify uncompromisingly with an all-embracing ideology that narrowly prescribes their attitudes, feelings, and practices. These ideologies are normative systems that define membership in groups that, to outsiders, appear relatively extremist—groups that are ethnocentric and xenophobic, homogeneous and intolerant of dissent and diversity, attitudinally and behaviorally narrowly focused and consensual, hierarchically structured internally, sharply delineated from other groups, and so forth. Such groups are often, at the least, undesirable, and at worst evil—for example, ultranationalist neo-Nazi groups, terrorist cells, and some religious cults. However, they can also be beneficial, a pro-social force for good—for example, some environmental groups, some religious orders, and some ethnolinguistic groups. In addition, not all totalist groups have all the attributes listed above. Finally, totalist groups have some of the attributes of active minorities, and therefore such groups may be a potent force for social change (e.g., Martin & Hewstone, 2001; Moscovici, 1980; Mugny, 1982). An understanding of the dynamics of membership in such groups is clearly an important conceptual challenge for social psychology and a practical challenge for society. In this chapter I explore the conditions under which people join totalist groups, make existing groups more totalist, or create new groups that have totalist attributes. The proposal is straightforward—extreme subjective uncertainty, particularly relating to self-conception, motivates people to define themselves, via selfcategorization and social identification processes, in terms of the highly consensual, distinctive, and clearly focused prototypes of high entitativity groups. This analysis is framed by the social identity perspective (e.g., Hogg, 2002; Hogg & Abrams, 1988; Tajfel & Turner, 1986; Turner, Hogg, Oakes, Reicher, &
UNCERTAINTY AND EXTREMISM
299
Wetherell, 1987), and develops the uncertainty reduction theory of social identity motivation (Hogg, 2000a, 2001a; Hogg & Mullin, 1999) to incorporate notions of entitativity (e.g., Hamilton & Sherman, 1996; Hamilton, Sherman, & Castelli, 2002; Sherman, Hamilton, & Lewis, 1999). The chapter opens by examining the general question of what is the relationship between uncertainty, identity, and totalism. I then briefly review key aspects of the social identity perspective and outline in detail the uncertainty reduction hypothesis. I extend the analysis with a conceptual discussion of prototypicality and entitativity and describe some relevant studies. The chapter closes with some concluding comments. UNCERTAINTY AND THE POSTMODERN PARADOX Postmodernity Contemporary Western society is often characterized as postmodern, in the sense that there is no absolute morality and few limitations on the sorts of people we can subjectively be or become—a world in which “reality” is what we construct it to be, and each person’s reality is as authentic as every other person’s reality. In the postmodern world one’s self-concept is largely individual and personal. It is a portable and infinitely malleable construction that can be reconfigured as we wish. Subjectively, there are few limits to self-conception and few absolutes that can guide us to the type of person we ought to be. In many ways this state of affairs is the natural consequence of a process of secularization, enlightenment, and industrialization that has been underway since the seventeenth century. For example, Baumeister (1987) paints a picture of medieval society in which social relations were fixed and stable, and were legitimized in religious terms. People’s lives and identity were clearly mapped out according to position in the social order—by visible ascribed attributes such as family membership, social rank, birth order, and place of birth. People were tightly locked into a matrix of prescribed group memberships and social relations. Over the past two to three hundred years, Western cultures have devel oped in the shadow of an economic system based on labor mobility. People have become units of production who are expected to move from places of low labor demand to places of high labor demand; they are expected to reorganize their lives, their relationships, and their self-concept around mobility, change, and transient relationships. Independence, separateness, and uniqueness have become more important than connectedness and the longterm maintenance of enduring relationships, these values have become enshrined as key features of postmodern society. The nineteenth-century sociopolitical theorist Tönnies (1887/1955) captured this temporal change in the sorts of self-defining relationships that people have, rather nicely, with the German terms Gemeinschaft and Gesellschaft (translated by the English words community and association). Sociologically oriented commentators have observed that the stable identities of a hundred years ago had, by the 1950s, been replaced by a more atomistic individual-oriented status society that “did not promote feeling of identification or collective involvement” (Nisbet, 1959, p. 17). This has produced a characteristic postmodern paradox that people with today’s less-structured self yearn for community and the collective affiliations of times past (e.g., Barber, 1995; Bashevkin, 1991; Dunn, 1998; Gergen, 1991; McNeil, 1986; Naisbitt, 1984); it is a yearning that may be reflected in contemporary religious fundamentalism, ethnic and racial revival, and the reemergence of nationalism and “new” racism (e.g., Barker, 1981; Billig, 1991; Hogg & Hornsey, 1998). People are striving to construct a more certain sense of self in an increasingly uncertain world.
300
THE PSYCHOLOGY OF GROUP PERCEPTION
Totalism and Extremism This analysis suggests that self-conceptual uncertainty, grounded in postmodernity, may play a role in motivating people to seek membership in totalist groups. A similar link between societal uncertainty and extremism has been proposed by Staub (1989), who suggests that extreme and enduring uncertainty, due to socioeconomic crises or natural disasters, may be a powerful motivational factor in the subsequent emergence of extremist groups that engage in group violence and genocide. In a similar vein, uncertainty has been proposed as a significant motivation for people to join extremist political parties (e.g., Billig, 1991; Hogg & Hornsey, 1998; Hogg & Reid, 1998) and “totalist” groups such as cults (e.g., Curtis & Curtis, 1993; Galanter, 1989). Marris (1996) places uncertainty reduction at the motivational core of intergroup relations. He argues that intergroup relations are a struggle to transfer uncertainty to other groups and so to construct a hierarchy of uncertainty with desirable high status groups characterized by low uncertainty. Widespread and extreme subjective uncertainty thus appears to play a role in the emergence and persistence of totalist groups and identities. It may play a role in ethnic, ethnolinguistic, and nationalist revivals (e.g., Serbian nationalism, Welsh, Basque, and francophone revivals in Wales, Spain, and Canada), cults and sects (e.g., Falun Gong), religious, political, and scientific orthodoxy (e.g., Taleban in Afghanistan, Amish in the United States, neo-Nazis in Germany), organizational cliques and “youth cultures” (e.g., hip-hop, rap), and so forth. The clear implication is that group membership, particularly membership in totalist groups, is functional in addressing people’s uncertainty resolution needs. The social identity perspective provides one way in which to understand this process. SOCIAL IDENTITY PERSPECTIVE The social identity perspective (e.g., Hogg & Abrams, 1988; Tajfel & Turner, 1986; Turner, Hogg, Oakes, Reicher, & Wetherell, 1987) has become increasingly central to social psychology (Abrams & Hogg, 1998; Hogg & Abrams, 1999; Moreland, Hogg, & Hains, 1994), and has recently been summarized in detail elsewhere (e.g., Abrams & Hogg, 2001; Hogg, 2001b, 2003). I provide only a brief overview of key features here. Overview of Key Features From the social identity perspective, a group exists psychologically when people share a self-conception in terms of the defining features of a self-inclusive social category. More specifically, this representation of the group is a prototype, a fuzzy set of features that captures ingroup similarities and intergroup differences regarding beliefs, attitudes, behaviors, and feelings. Prototypes are configured, according to the principle of metacontrast, to maximize the ratio of intergroup differences to intragroup differences and thus to render the categories as distinct entities. A key insight of the social identity perspective is that the basis of perception, attitudes, feelings, behavior, and self-conception is contextually fluid. Self-conception can vary from being entirely based on idiosyncratic personal attributes and the unique properties of a specific interpersonal relationship to being entirely based on a shared representation of “us” defined in terms of an ingroup prototype. In the latter case the situation represents a group situation and perceptions, attitudes, feelings and behavior acquire the familiar characteristics of inter- and intragroup behaviors—conformity, normative behavior, solidarity, cohesion, stereotyping, ethnocentrism, intergroup discrimination, ingroup favoritism, and so forth. Put another way, the more an aggregate of people is a salient basis for self-definition as a group member, then
UNCERTAINTY AND EXTREMISM
301
the more strongly is self-definition, perception, cognition, affect, and behavior based upon prototypicality. When group membership is the salient basis of self-conception, people, including the self, are represented and treated in terms of the relevant ingroup- or outgroup-defining prototype. Self-categorization depersonalizes the self in terms of the ingroup prototype (producing self-stereotyping, conformity, normative behavior, social attraction, social identification, and so forth), and it depersonalizes perception of others so that they are viewed and treated as more or less exact matches to the relevant prototype. Prototypicality is the yardstick of life in salient groups. Because groups define the self, the social value or status of a group becomes the social value or status of the self. Intergroup relations become, therefore, a struggle for evaluatively positive distinctiveness for one’s own group relative to other groups. The struggle for positive distinctiveness is, in turn, underpinned by a self-enhancement motive and a striving for positive self-esteem. The strategies that groups and their members adopt to manage positive distinctiveness and self-enhancement are influenced by people’s beliefs about the nature of relations between groups—the legitimacy and stability of status relations, the permeability of intergroup boundaries, and the possibility of an alternative social order. Self-Enhancement and Social Identity The principal motivation underlying social identity processes is self-enhancement, reflected in the pursuit of positive social identity grounded in positive intergroup distinctiveness. In intergroup contexts people strive to make ingroup-favoring social comparisons that secure evaluatively positive distinctiveness for the ingroup and thus for one’s social identity (Tajfel, 1972; Turner, 1975; see Hogg, 2000b). The underlying motivation is for self-enhancement or self-esteem (e.g., Turner, 1982; Turner, Brown, & Tajfel, 1979; also see Abrams & Hogg, 1988). Research suggests that self-enhancement certainly is a very basic human motivation (e.g., Baumeister & Leary, 1995; Sedikides & Strube, 1997), but the details of its exact relationship to social identity processes are less clear (e.g., Abrams & Hogg, 1988; Hogg & Abrams, 1990, 1993; Long & Spears, 1997; Rubin & Hewstone, 1998). For example, although lower self-esteem sometimes motivates group identification, it sometimes does not. Other research suggests that self-esteem does not motivate behavior at all, but rather acts as an internal monitor of successful social connectedness (e.g., Leary, Tambor, Terdal, & Downs, 1995; also see Baumeister & Leary, 1995). Research on the self-esteem hypothesis distinguishes between individual and group membership based self-esteem and suggests that the relationship between self-esteem and group behavior may be affected by other variables such as the extremity of self-esteem, the degree to which people identify with the group, and the extent to which groups and their members may feel under threat. Self-enhancement clearly plays a motivational role in social identity processes, for people would rather belong to positively valued groups, they strive to make their groups positively valued, and they are adept at personally avoiding or denying undesirable evaluative consequences of membership in stigmatized groups (e.g., Crocker & Major, 1989; Crocker, Major, & Steele, 1998). However, groups also provide a clear sense of who one is and of how one should react to and what one should expect from others. Group membership defines the self and prescribes one’s attributes and one’s relationships to ingroup and outgroup others. Group membership and social identity provide one with a degree of self-conceptual certainty. Presumably, then, social identity processes may be fully or partially motivated by uncertainty reduction concerns.
302
THE PSYCHOLOGY OF GROUP PERCEPTION
UNCERTAINTY REDUCTION HYPOTHESIS The idea that uncertainty reduction may motivate social identity processes was briefly introduced by Hogg and Abrams (1993) and fully developed by Hogg (2000a, 2001a, 2001c; Hogg & Mullin, 1999). The core idea is that subjective uncertainty reduction is a fundamental human motivation that is extremely well satisfied by social identification—more precisely, by prototype-based depersonalization based on selfcategorization. Motivational Status of Uncertainty Reduction People have a fundamental need to feel certain about their world and their place within it. Subjective certainty renders existence meaningful and thus gives one confidence in how to behave and what to expect from the physical and social environment within which one finds oneself. Uncertainty about one’s attitudes, beliefs, feelings and perceptions, as well as about oneself and other people, is aversive (e.g., Lopes, 1987; Sorrentino & Roney, 1999) because it is ultimately associated with reduced control over one’s life. Thus, uncertainty motivates behavior that reduces subjective uncertainty. The search for certainty is nearer Bartlett’s (1932) notion of a search after meaning than James’s (1890) argument that people strive to simplify their experiences. This is not to say that people don’t try to simplify their experiences, but that any simplification that occurs is associated with a more fundamental search for certainty. An overly simplified worldview would be so crudely textured that it would be functionally meaningless, and it would therefore not provide subjective certainty. People do not strive for certainty about all aspects of life, but only those that are subjectively important. Consistent with the cognitive miser and motivated tactician models of social cognition (Fiske &: Taylor, 1991; S.E. Taylor, 1998), people would only be expected to expend cognitive effort reducing uncertainty when the reduction of uncertainty really matters. Subjective uncertainty is produced by contextual factors that challenge peoples certainty about their cognitions, perceptions, feelings, and behaviors, and ultimately, certainty about and confidence in their sense of self. Self-certainty can be undermined because the self is the critical organizing principle, referent point, or integrative framework for diverse perceptions, feelings, and behaviors. The locus of uncertainty is to be found overwhelmingly in the social context, and therefore everyone is prone to uncertainty. However, biographical factors may have some influence on people’s orientation towards uncertainty and the reduction of uncertainty. Epistemic motives related to uncertainty have a high profile in social psychology. For example, they make an appearance in the recent explosion of research on the self and on self-motives (e.g., Banaji & Prentice, 1994; Sedikides & Strube, 1995; see Leary & Tangney, 2002), which considers the quest for selfdefinition to be a persistent and central feature of human existence (e.g., Gaertner, Sedikides, & Graetz, 1999). Very early on, Festinger (1954a, 1954b) believed that there is a “motivation to know that one’s opinions are correct and to know precisely what one is and is not capable of doing” (Festinger, 1954b, p. 217). Knowing one is correct (cf. uncertainty reduction) is a critical human motivation that drives people to make interpersonal social comparisons with individual others when nonsocial means are unavailable. Communication researchers believe that people communicate to reduce uncertainty and that effective communication itself requires interpersonal certainty (e.g., C.R.Berger, 1987; C.R.Berger & Bradac, 1982; Gudykunst & Shapiro, 1996; Gudykunst, Yang, & Nishida, 1985). Self-presentation researchers believe that although impression management is an important motive in self-presentation (e.g., Baumeister, 1982), another important motive is self-construction; we present ourselves publicly to others in order to confirm who we are (Tice, 1992). Self-construction represents internalized self-conceptual change on the basis of
UNCERTAINTY AND EXTREMISM
303
other people’s commitment to the publicly presented behavior (Schlenker, Dlugolecki, & Doherty, 1994; Tice, 1992), and the self-presentation process that produces this outcome need not be deliberate, it can be automatic (Paulhus, 1993; Tice, Butler, Muraven, & Stillwell, 1995). Many social psychologists study individual differences in uncertainty-related motives: for example, Adorno, Frenkel-Brunswik, Levinson, and Sanford’s (1950) description of the authoritarian personality, Rokeach’s (1948, 1960) notion of a dogmatic or closed-minded personality, Sorrentino’s distinction between uncertainty-oriented and certainty-oriented people (e.g., Sorrentino & Roney, 1999), and Kruglanski and others’ research on need for structure and the need for closure and the fear of invalidity (e.g., Kruglanski, 1989; Kruglanski & Webster, 1996; M.N. Thompson & Zanna, 1995; Neuberg & Newson, 1993; Webster & Kruglanski, 1994). There are also other constructs that identify individual differences in the complexity and number of explanations people have of other people (attributional complexity: Fletcher, Danilovics, Fernandez, Peterson, & Reeder, 1986), individual differences in how much people like to think deeply about things (need for cognition: Cacioppo & Petty, 1982), and individual differences in the complexity of people’s cognitive processes and representations (cognitive complexity: Crockett, 1965). People also differ in selfconcept clarity (the extent to which self-beliefs are clearly and confidently defined, internally consistent, and stable: e.g., J.D.Campbell, 1990; J.D.Campbell, Trapnell, Heine, Katz, Lavalle, & Lehman, 1996); selfcomplexity (the number of different or independent dimensions that underlie self-conception: e.g., Linville, 1987); and compartmentalization of the self (e.g., Showers, 1992). From a social identity perspective, personality and individual difference conceptualizations of uncertainty should be treated cautiously. It is certainly not my intention to trace group behaviors such as discrimination and prejudice to personality (cf., Billig, 1976). Rather, social contextual factors influence uncertainty, the resolution of uncertainty, and the way in which such resolution is expressed. If predispositions have a role to play, it is a relatively minor role that is tightly constrained by social context. Individual differences may be better understood as a function of enduring social contexts rather than personality differences; uncertain times produce uncertain people. Reducing Subjective Uncertainty by Self-Categorization People can probably reduce subjective uncertainty in many different ways. For example, according to social comparison theory, people first make direct physical reality checks. If physical reality checks are impossible, people make interpersonal comparisons with interpersonally similar others. The idea I propose here is that another very powerful way to reduce subjective uncertainty is through self-categorization, and indeed that uncertainty reduction is a core motive for social identity processes, perhaps a more fundamental motive than self-enhancement. The process of self-categorization that is responsible for social identification and group membershipbased behaviors itself reduces subjective uncertainty. Social categorization constructs contextually relevant ingroup and outgroup prototypes on the basis of the principle of metacontrast. Such prototypes capture meaningful similarities within and differences between groups, and the categorization process further accentuates prototypical similarities within groups and differences between groups. The perceptual field is rendered more clear and meaningful and less complex and perplexing; it is brought clearly into focus in subjectively meaningful ways. Social categorization of the self (self-categorization) depersonalizes the self in terms of the ingroup prototype. The self is contextually transformed so that self-conceptualization, attitudes, feelings, and behaviors are governed by the ingroup prototype, a prototype that is both descriptive and prescriptive. This process also provides an ingroup social comparative context containing similar others
304
THE PSYCHOLOGY OF GROUP PERCEPTION
(physically or cognitively present) who all appear to validate one’s self-concept and associated cognitions and be haviors. Thus, uncertainty about self-conceptualization, attitudes, feelings, perceptions, and so forth is reduced. The process of self-categorization reduces uncertainty, and, in a more enduring sense, people may join (identify with) groups because they reduce uncertainty. Not all uncertainty motivates uncertainty reduction. The focus of uncertainty needs to be subjectively important in that context (something that matters in one context may not matter in another context) for uncertainty to motivate people to try to reduce uncertainty. Furthermore, it is very likely that one class of contextually important uncertainties is those that ultimately relate to self-conception; they represent uncertainty about things that define the self, and therefore about one’s cognitions, feelings, and actions. Resolution of these important uncertainties may almost necessarily imply self-conceptual realignment rather than, or in addition to, simply acquiring knowledge. The self-categorization process is ideally suited to this form of uncertainty reduction. One important caveat is that although self-relevant uncertainty may be resolved by identification with “any” cognitively accessible self-inclusive social category, people will generally show preferential identification with categories whose prototypical features seem most relevant to the dimensions on which uncertainty is being experienced. For example, if I am uncertain about how to define myself with reference to my attitudes about literary genres, I am more likely to identify with a literary movement than with a football fan club—the prototypical properties of the former speak more to literary opinions than do the prototypical properties of the latter. Group prototypes play a pivotal role in uncertainty reduction (see below). Uncertainty Reduction and Self-Enhancement Uncertainty reduction and self-enhancement are both motives for social identity processes. People strive to identify with groups that resolve self-conceptual and more general uncertainty and with groups that contribute positively to self-evaluation. There is probably a dynamic relationship between these two motives (e.g., Reid & Hogg, 2003). Threats to identity valence will arouse the self-enhancement motive, and threats to identity distinctiveness the uncertainty reduction motive. Depending on which motive has been aroused (though, of course, many contexts may actually arouse both motives), one would expect different social identity-related behaviors. For example, under social identity valence threat people would engage in behaviors designed to repair or promote the status of their group, or they would try to identify with groups of higher status—these strategies are those described by the macrosocial aspect of social identity theory (e.g., Tajfel & Turner, 1986; also see Ellemers, 1993; Hogg & Abrams, 1988). Under social identity distinctiveness threat people would engage in behaviors designed to repair or promote the distinctiveness or entitativity of their group, or they would try to identify with groups of greater entitativity. One manifestation of these behaviors might be consolidation and justification of the existing social order—a phenomenon related to that described by system justification theory (e.g., Jost, 1995; Jost & Banaji, 1994). Another manifestation might be a tendency to reject all prototypically marginal ingroup members equally, even those whose attributes contribute positively to the valence of the group (e.g., Hogg, Fielding, & Darley, in press); typically, rejection of ingroup deviants would be expected to fall more heavily on negative deviants, so called black sheep (e.g., Marques & Páez, 1994; also see Marques, Abrams, Páez, & Hogg, 2001). Relatedly, under distinctiveness threat, people, particularly those who are highly identified with the group, might adopt very strict criteria for admission of any new members to the group—an overexclusion effect (e.g., Leyens & Yzerbyt, 1992; Yzerbyt, Leyens & Bellour, 1995).
UNCERTAINTY AND EXTREMISM
305
THE ROLE OF PROTOTYPICALITY AND ENTITATIVITY IN SELFCATEGORIZATION-BASED UNCERTAINTY REDUCTION The mechanism of uncertainty reduction is assimilation of the self to the ingroup prototype. It is the prototype that defines one’s self-concept and prescribes one’s perceptions, attitudes, feelings, and behaviors as a group member. The prototype provides focus for and reduces uncertainty about self-conception, cognition, affect, and action. Self-categorization depersonalizes the self in terms of the properties of the ingroup prototype—it places the ingroup prototype in the driver’s seat. Therefore, the properties of the prototype, the ingroup defined by the prototype, and the wider intergroup context that generates the prototype all have important effects on self-categorization-based uncertainty reduction. Self-categorization is more effective in reducing uncertainty if people have a concentrated, clearly focused, relatively simple, and relatively unambiguous prototype. Such prototypes would typically have a small number of attributes that are unambiguous, highly correlated, clearly prescriptive, and highly diagnostic of ingroup membership. It is possible that under some circumstances the prototype may take on the properties of an exemplar. For instance, nations experiencing turmoil and uncertainty often present themselves in terms of images of their “great leader” (e.g., Lenin, Ho Chi Minh, Mao Zedong, Ayatollah Homeini, Saddam Hussein)—images that represent the essence of the nation. If the prototype is nonconsensual and is prescriptively unclear then it is only of limited use in regulating cognition and behavior in unambiguous ways that confer subjective certainty. Indeed, if people cannot form a prototype at all, then self-categorization does not reduce uncertainty, and people will pursue other means or other self-categorizations to reduce uncertainty. However, people can be creative in forming prototypes of groups that seem quite bereft of con tent; for example in minimal groups people may use what they know about themselves to generate a prototype of the group (Cadinu & Rothbart, 1996; also see Tropp & Wright, 2001). Generally, the prototype should, as described above, also be relevant to the dimension of uncertainty. It should embrace properties that address the specific dimension of uncertainty, or it should be broadly enough relevant to the self-concept to inform the dimension of uncertainty. To reduce uncertainty, people will identify more readily with groups that are relevant to aspects of self-conceptualization that are challenged by the specific contextual uncertainty that is being experienced. Typically, clearly focused prototypes of the type discussed above are best furnished by relatively homogeneous, consensual groups that are high in entitativity (e.g., Campbell, 1958; also see Brewer & Harasty, 1996; Hamilton & Sherman, 1996; Hamilton, Sherman, & Castelli, 2003; Hamilton, Sherman, & Lickel, 1998; S.J.Sherman, Hamilton, & Lewis, 1999). Groups certainly vary in entitativity (Lickel, Hamilton, Wieczorkowska, Lewis, and Sherman, 2000), with high entitativity groups having sharp boundaries that map out the contours of quite distinct prototypes. Furthermore, such groups are located in a crystallized intergroup context where intergroup relations are stable and clearly delineated. Indeed, in seeking uncertainty reduction people probably pay as much attention to clarity of social structural differentiation among groups as to ingroup prototype clarity itself; although the former is a reliable cue to prototype clarity, the latter is probably the most direct source of self-definitional information (Yzerbyt, Castano, Leyens, & Paladrino, 2000). In any event, according to the metacontrast principle prototype formation rests on both intergroup differentiation and intragroup homogenization. Prototypes are dynamic representations, and due to their critical role in uncertainty reduction, it is very likely that under conditions of uncertainty people pay very close attention to prototypicality. Under these conditions prototype-relevant information will be processed systematically via a central processing route, rather than heuristically via a peripheral route (e.g., Eagly & Chaiken, 1993; Petty & Wegener, 1998).
306
THE PSYCHOLOGY OF GROUP PERCEPTION
People will do cognitive work on prototypes in order to construct them in ways that maximize their uncertainty reduction properties. Thus, systematic processing of prototype information will be more evident when individuals or groups confront greater uncertainty, for example, minority groups, or majority groups faced by growing challenges to their hegemony. TRANSFORMING UNCERTAINTY INTO EXTREMISM Subjective uncertainty reduction is an important human motive that is well satisfied by self-categorization in terms of a group that has a clear prototype. The greater the self-conceptual uncertainty, the greater the need to identify with groups with very clear prototypes. Such prototypes are provided by high entitativity groups —groups characterized by great similarity, proximity, boundedness, common fate, organization, structure, and so forth. One implication of the role of prototype clarity in uncertainty reduction is that extreme, and possibly chronic, uncertainty about subjectively critical matters that impact the self-concept may cause people to identify with extreme groups that are highly orthodox (very clear prototypes) and very distinct from other groups. These groups may also have hierarchical internal power and leadership structures (Hogg, 2001d, 2001e; Hogg & Reid, 2001; also see Hogg & van Knippenberg, 2003). People may create new groups that have these properties, join existing groups that have these properties, or make groups they already belong to more extreme and more orthodox. Prototype clarity, group entitativity, and socio-structural stability are further enhanced by ideological and attribution processes that transform perceptions such that social attributes of groups are viewed as immutable and natural underlying “essences” of kinds of individuals (Haslam, Rothschild, & Ernst, 2000; Yzerbyt, Corneille, & Estrada, 2001). This accentuates the immutability, distinctiveness, and representational clarity of such groups. People striving to construct a more certain sense of self in society, invoke categories that can more readily be represented in psychological essentialist terms (Medin & Ortony, 1989; Miller & Prentice, 1999). The psychology of extremism may thus hinge on uncertainty reduction through self-categorization. The greater the uncertainty, the greater the motivation to structure the self-concept in terms of the prototypical properties of high entitativity groups—groups that are orthodox, distinctive, hierarchically structured, consensual, intolerant of internal dissent and diversity, and highly xenophobic and ethnocentric. One important caveat is that identification with a high entitativity group in order to resolve extreme uncertainty does not necessarily mean that one never again will experience self-conceptual uncertainty. Conditions may arise that threaten entitativity in various ways, and from the relative security of totalism people may once again feel “safe” to focus on ways in which fellow group members challenge the prototypical integrity of the group. For example, Corneille and Judd (1999) found that although group identification produced the usual accentuation effects on prototypical dimensions, it actually increased perceived intragroup variability on nonprototypical dimensions. STUDIES OF UNCERTAINTY AND IDENTIFICATION WITH HIGH ENTITATIVITY GROUPS The core uncertainty reduction hypothesis has been tested in a series of laboratory, mainly minimal or quasiminimal group, experiments (Grieve & Hogg, 1999; Hogg & Grieve, 1999; Hogg & Mullin, 2000; Mullin & Hogg, 1998,1999; Reid & Hogg, 2003) that have been summarized elsewhere (Hogg, 2000a, 2001a; Hogg & Mullin, 1999). The general paradigm is one in which (a) contextual subjective uncertainty or type of
UNCERTAINTY AND EXTREMISM
307
FIGURE 21.1. Ingroup bias (three-item scale, −12 to +12, =.63) as a function of categorization and uncertainty, F (1, 104)=6.21, p<.05, 2=.06 (Grieve & Hogg, 1999, Experiment 2). The mean of 3.43 is significantly different from zero, and is different from all other means.
uncertainty is manipulated, (b) participants are explicitly categorized or not (or category salience is manipulated), or (c) prototypical properties of the category are manipulated. The key dependent measures are self-report measures of group identification and behavioral measures of resource allocations. To illustrate the general nature of these studies, let me just describe one. Grieve and Hogg (1999, Experiment 2) conducted a standard minimal group experiment, presented as a study of decision making, in which categorization and uncertainty were manipulated in a 2×2 between-participants design. Participants had their uncertainty raised or lowered by performing a judgment task in which the correct judgments were ambiguous or obvious; they had to write what they thought was happening in a series of ambiguous pictures (from the Thematic Apperception Test) or a series of unambiguous pictures (e.g., a picture of a footballer kicking a goal). Checks revealed this manipulation to be highly effective. They were then explicitly randomly categorized into X and Y groups, or they were not categorized. This was done under normal minimal group conditions—complete anonymity, no interaction, no past or future for the group, and so forth. They then completed the standard minimal group point-allocation matrices, and went on to answer self-report questions monitoring group identification. On a composite three-item measure of ingroup bias ( = 0.63), there was significant ingroup bias only among participants who were categorized under elevated uncertainty (see Figure 21.1). These people also identified (on a composite 10-item identification scale, =0. 91) more strongly with the group (see Figure 21.2). Together the studies cited above show quite robustly that (a) people who are explicitly categorized only self-categorize and thus identify with the category, express group attitudes, and engage in group behaviors when they are categorized under conditions of subjective uncertainty; (b) the effect occurs largely irrespective of how uncertainty is caused; but (c) it is most pronounced when the dimension of subjective uncertainty is subjectively important and when the social category is relevant to the dimension of uncertainty. There is also some evidence that uncertainty still motivates identification even after people have had an opportunity to address any reduction in self-esteem that may have been inflicted by the uncertainty manipulation. An interesting and relevant finding from a different paradigm is that the perceptual accentuation effect (e.g., Tajfel, 1959), in which people perceptually accentuate ingroup similarities and intergroup differences
308
THE PSYCHOLOGY OF GROUP PERCEPTION
FIGURE 21.2. Group identification (10-item scale, 1 to 9, =.91) as a function of categorization and uncertainty, F (1, 104)=7.61, p<.01, 2=.07 (Grieve & Hogg, 1999, Experiment 2). The mean of 5.93 is significantly different from all other means. The mean of 4.67 is also significantly different from all other means.
on dimensions believed to be systematically associated with the social categorization, may only occur under uncertainty. Corneille, Klein, Lambert, and Judd (2002; also see Lambert, Hogg, Klein, & Azzi, 2002) show how Tajfel & Wilkes’s (1963) classic finding relating to the judgment of line lengths occurred only because the participants in that study, who were British and thus familiar with inches, were asked to make judgments in centimeters—the participants were uncertain. Where participants are more familiar with the judgment dimension, and therefore less uncertain, the effect weakens. Corneille and colleagues found a stronger accentuation effect among Americans judging in centimeters and Belgians judging in inches than vice versa. Direct studies of the role of the prototype and of entitativity in self-categorization mediated uncertainty reduction are less plentiful. There is some support, from two minimal group studies, for the idea that people who are uncertain are more likely to identify with a more than less homogeneous ingroup (Jetten, Hogg, & Mullin, 2000). In Experiment 1 participants were made more or less uncertain, using Mullin and Hogg’s (1999) method, in which participants gave their opinions on important lifestyle issues and were told that there was an objectively correct set of opinions (high uncertainty) or that everyone’s opinions were equally correct (low uncertainty). They were then categorized on the basis of a dot-estimation task and given information that their group was homogenous or heterogenous regarding members’ dot estimations. Multiitem checks confirmed the effectiveness of the manipulations. The main index of identification was a threeitem scale measuring the difference between how well the ingroup matched the ingroup stereotype and the outgroup matched the outgroup stereotype. As predicted, participants who were uncertain and had been categorized as members of an homogeneous group showed greater intergroup differentiation on stereotypical dimensions than did other participants (see Figure 21.3. Experiment 2 was a virtual replication of Experiment 1, except that participants were given the opportunity to obtain information about fellow group members’ decisions on the uncertainty task (an attitude task) and change their own decisions accordingly. There was an almost significant tendency for participants who were uncertain and had been categorized as members of a homogeneous group to request more information and to make more changes in their decisions. Minimal groups do not have rich prototypes, so a naturalistic context was sought to investigate the idea that uncertainty, particularly self-conceptual uncertainty, may motivate people to join groups, particularly high entitativity groups with clearly focused prototypes. Students arriving at university can experience uncertainty, including uncertainty about how they will fit in and what sort of identity they will have. They
UNCERTAINTY AND EXTREMISM
309
FIGURE 21.3. Stereotypical differentiation (three-item scale, −8 to +8, =.78) as a function of uncertainty and group variability, F (1, 52)=6.93, p<.01, 2=.12 (Jetten, Hogg, & Mullin, 2000, Experiment 1). The mean of .17 differs from the means of 1.93 and 1.43. The three larger means do not differ from one another.
can also join university clubs and societies that vary in terms of entitativity and prototype clarity. We conducted two questionnaire studies (Hogg & Sussman, 1999), one with new psychology students at the University of Queensland and the other with new students at Princeton University. In each case we had about 200 respondents. At the start of the semester we measured demographic and background variables, as well as the focal variables of (a) self-conceptual uncertainty, (b) intention to join clubs and societies, and (c) perception of how distinctive they felt the clubs they were intending to join were relative to other clubs (a measure of entitativity). Using hierarchical multiple regression to control for demographic and background variables, we found, for both samples, a positive association between uncertainty and the intention to join clubs, to join more clubs, and to join clubs that were distinctive. With the Princeton sample we administered a follow-up questionnaire near the end of the academic year to remeasure uncertainty and to measure identification with clubs and societies they had joined. We found that uncertainty at time-one predicted identification with clubs at time-two, and that at time-two identification was associated with a reduction of uncertainty from time-one. In line with the uncertainty reduction hypothesis, these studies suggest that the more self-conceptually uncertain people are, the more they want to join groups, in particular, high entitativity groups with clear prototypes, and the more strongly they identify with these groups the greater the reduction in uncertainty. One final study took advantage of political extremism in Australia (Hogg & Reid, 1998). At the very end of the 1990s an ultraconservative political party, called One Nation, emerged. It had all the properties of extremist right-wing parties that are xenophobic, orthodox, racist, nationalistic, anti-immigration, and antiintellectual (see Rapley, 1998). It was also a very high entitativity group, with a very clear and simple prototype. Hogg and Reid (1998) administered a questionnaire to about 200 university students at a large urban and a smaller regional university in the southeast of the state of Queensland. We asked respondents a wide range of questions relating to their feelings of uncertainty about their own economic, employment, and interpersonal future and about the economic and cultural future of Australia. We also focused on uncertainty about what sort of a person they were. Since Australian university students typically do not endorse One Nation, we framed questions about identification with One Nation rather carefully—focusing on endorsement of One Nation’s ideology and worldview rather than on identifying with One Nation. Regression analyses revealed, as predicted, that increasing subjective uncertainty was associated with increasing endorsement of One Nation’s worldview.
310
THE PSYCHOLOGY OF GROUP PERCEPTION
Further, indirect, support for the idea that highly uncertain people may seek out high entitativity groups with which to identify comes from a series of four studies conducted by Castano, Yzerbyt, and Bourguignon, summarized by Yzerbyt, Castano, Leyens, and Paladrino (2000, pp. 282–285). European participants’ attitudes toward and identification with the European Union (EU) were measured. They were then confronted with information about the EU that represented it as possessing high entitativity or low entitativity, and were asked to indicate again how strongly they identified with the EU. Very broadly, Castano and associates found that people holding moderate attitudes toward the EU (as opposed to people with strong pro or anti attitudes) showed a contextual increase in identification, but most notably when presented with a high entitativity representation of the EU. Although uncertainty was not manipulated or measured, it is plausible to conjecture that moderate attitudes toward the EU indicate a greater degree of uncertainty than do strong pro or anti attitudes. CONCLUDING COMMENTS My goal in this chapter has been to explore the role of social identity processes in extremism. Specifically, I have argued that the reduction of subjective uncertainty is a key motivation for social identification and that the process of social identification is a very effective mechanism for reducing subjective uncertainty. The underlying mechanism is self-categorization, and thus depersonalization, in terms of the ingroup prototype. Self-categorization transforms self-conception and effectively internalizes the group prototype as the standard for one’s perceptions, attitudes, feelings, and actions. Prototypes describe and prescribe cognition, affect, and action. Furthermore, the social comparative context of depersonalization usually provides social validation of one’s self-conception, perceptions, feelings, attitudes, and behaviors. It is the prototype that reduces uncertainty, and thus concentrated, clearly focused, relatively simple, and relatively unambiguous prototypes are more effective. Such prototypes have a small number of attributes that are unambiguous, highly correlated, clearly prescriptive, and highly diagnostic of ingroup membership. Prototypes such as this define high entitativity groups—groups characterized by similarity, proximity, boundedness, common fate, tight organization, clear structure, and so forth. In extreme cases such groups could be thought of as “totalist” or extremist—groups which are highly orthodox (very clear prototypes) and very distinct from other groups, and which may have hierarchical internal power and leadership structures. The final step in the argument is that extreme uncertainty will cause people to identify with such groups. People may create new groups that have these properties, join existing groups that have these properties, or make groups they already belong to more extreme and more orthodox. This analysis suggests a social cognitive mechanism that may account for the relatively well-documented observation that extremism often arises out of uncertainty. The chapter opened with a discussion of the postmodern paradox and some other macrosocial commentaries on uncertainty and extremism. I think the analysis provided in this chapter furnishes a social psychological mechanism to help account for these phenomena. The empirical evidence for this analysis comes from more than 30 years of research on social identity processes, a recent program of research on the uncertainty reduction hypothesis, and recent research on entitativity, essentialism, and group processes. Direct tests of the specific proposition that tightly focused consensual prototypes resolve uncertainty, and that very high entitativity groups possess such prototypes is less substantial, but is the focus of an ongoing research program.
22 Entitativity and Social Integration Managing Beliefs about the Reality of Groups RICHARD L.MORELAND JAMIE G.McMINN Department of Psychology, University of Pittsburgh
At or near the heart of this book is an intriguing paper about groups written more than 40 years ago by Campbell (1958). For many years, that paper had little impact on social psychologists who study groups. In fact, simply referring to the paper was long viewed as evidence of scholarship among some group researchers. But in recent years the papers impact has increased considerably (see, for example, Dasgupta, Banaji, & Abelson, 1999; Lickel, Hamilton, Wieczorkowska, Lewis, Sherman, & Uhles, 2000; McConnell, Sherman, & Hamilton, 1997; Pickett, 2001). Why? Maybe because the paper has been interpreted as an analysis of how groups are perceived, which links it to a powerful trend in recent research on groups (Moreland, Hogg, & Hains, 1994), namely the use of social cognition theories and methods to study group behavior (see Fiske & Goodwin, 1996). Although it is encouraging that Campbell’s (1958) paper has become more influential, we are concerned about its interpretation and the kinds of research that it has produced. A careful reading of the paper suggests that it was not only about how groups are viewed, but also about the reality of groups themselves. Research in this area is thus inadequate, because it focuses on just one aspect of a complex construct. ENTITATIVITY VERSUS SOCIAL INTEGRATION Campbell’s (1958) paper was about group “entitativity,” which he defined as “the degree of having the nature of an entity, of having real existence” (p. 17). Note that this definition, as stated, emphasizes whether groups are real, not whether they seem real. In fact, Campbell began his paper with a brief discussion of the debate among some scholars about whether a group is more like an organism or a machine.1 But Campbell quickly set aside that issue to consider instead how people decide whether a set of individuals is a group at all. In that regard, he identified and discussed several cues, including common fate, similarity, proximity,
312
THE PSYCHOLOGY OF GROUP PERCEPTION
and prägnanz (form). Several individuals, Campbell said, will seem more like a group when their outcomes are interdependent, or when they are similar to one another, remain close to one another, or take the form (e.g., a line or a circle) of a group. It is important to realize, however, that Campbell saw the use of such cues, along with the decisions they produce, as just the first step in a two-step process (p. 15). First, people decide whether a set of individuals is a group. Then, if a group is detected, they go on to consider what kind of group (organismic or mechanical) it might be. It seems ironic that Campbell was worried about how little psychologists knew about the first step in this process (p. 16). Today, he might well complain that the second step has been ignored. Although Campbell’s (1958) paper is difficult to interpret at times, it contains several other clues that entitativity may mean more than just beliefs about the reality of groups.2 First, the examples Campbell used to clarify the use of entitativity cues (e.g., stones, teacups) were all objects that are clearly real, whether they seem real or not. Their existence, in other words, does not depend on anyone’s beliefs about their reality. Second, Campbell made an evolutionary argument for the development and use of entitativity cues. If groups were not real, he argued, then methods for their detection would not have evolved. The fact that people can and do “see” social groups thus implies that such groups can be real (p. 23). Finally, Campbell frequently mentioned perceptual illusions, suggesting that a real group can be overlooked, or a “group” that is not real can be “detected” by mistake. If our analysis of Campbell’s (1958) paper is correct, then entitativity means more than just whether groups seem real or not. Entitativity also refers to whether groups are real or not. This terminological situation is awkward. It would probably be best, in our opinion, to leave “entitativity” alone, as a term referring to the perceived reality of groups. Most researchers have used the term in that way, so it would be difficult to change its meaning at this point. Another word or phrase is thus needed, one referring to the reality of groups, or to put it rather oddly, their “groupiness.” Several theorists have analyzed this aspect of groups and identified some of the variables associated with it (see, for example, Borgatta & Cottrell, 1955; Cattell, 1948; McGrath, 1984; Sandelands & St. Clair, 1993). We focus here on the concept of “social integration,” which Moreland (1987) proposed as the key to group formation. Social integration is the degree to which a set of people acts, thinks, and feels like an individual. According to Moreland (1987), it can take environmental, behavioral, affective, and cognitive forms. Environmental integration involves such variables as the amount of time people spend together and their shared use of space. Behavioral integration involves needs (e.g., survival, information, belongingness) that require social interaction to be met, or are met more easily through interaction. Cognitive integration and affective integration involve interdependent thoughts and feelings, respectively. Although shared thoughts and feelings can reflect interdependence among group members, interdependence sometimes takes other forms as well. Consider, for example, a group that contains two cliques, each disliking the other because they take opposite positions on some important issue. Such a group has high levels of both cognitive and affective integration, because the thoughts and feelings of each clique are determined by those of the other clique. Moreland (1987) argued that social integration is continuous, not categorical. Instead of simply categorizing a group as real or not real, its level of social integration thus ought to be assessed. Groups can vary in their reality. Some groups are more real than others, and the same group can become more or less real as time passes. This may be difficult to grasp. Social psychologists still disagree about whether groups can be real or not, so the claim that their reality can vary is likely to startle some of them. We make sense of all this by relying on metaphors. As a group’s level of social integration rises, we thus think of it as being “healthier” or “stronger” or more “clear” and “focused.”
ENTITATIVITY AND SOCIAL INTEGRATION
313
FIGURE 22.1. Entitativity and social integration in groups (cf. Hamilton, Sherman, & Lickel, 1998). X1–X5 are cues that affect how real a group seems to people; Y1–Y5 are signs of the group’s social integration.
According to Campbell (1958), entitativity is also continuous, rather than categorical. As Hamilton, Sherman, and Lickel (1998) have noted, this creates a complex situation (see Figure 22.1). Social integration and entitativity are distinct aspects of groups, each with its own set of related cues. One interesting issue is how much those cues overlap. Some social integration cues are similar to those associated with entitativity. Proximity is a cue for environmental integration, for example; common fate is a cue for behavioral integration, and so on. But social integration and entitativity cues can also differ. Consider the group we described earlier, the one containing two conflicting cliques. Although that group had a high level of social integration, its entitativity level would probably be low.3 Another issue, of greater interest to us, is how social integration and entitativity relate to one another. These are distinct concepts, in principle, but they are probably correlated in practice. That correlation is positive, and natu rally it runs both ways. Groups that are more real are likely to seem more real, and groups that seem more real are likely to become more real, through such processes as labeling and reflection (members are called a group and treated like a group by others; see Lewis, 1973; Ryder, Kafka, & Olsen, 1971). The fact that social integration and entitativity can diverge, however, means that mistakes are possible, and if mistakes can be made, then people may try to deceive one another about the reality of groups, using various impression management tactics. MANAGING IMPRESSIONS ABOUT GROUPS There is, of course, much theory and research in social psychology on the process of impression management (see Leary, 1995; Schlenker & Weigold, 1992). Several motives for managing impressions have been identified. The most powerful of these seems to be social desirability, usually in the service of personal gain. Many of the tactics that people use to manage impressions, some simple and others more complex and subtle, have also been studied. Most of the research in this area, however, focuses on how
314
THE PSYCHOLOGY OF GROUP PERCEPTION
someone tries to control how he or she is viewed by others. The typical focus of such research, in other words, is impression management on one’s own behalf. The main exception to this trend is a small set of studies on attempts by one person to manage the impression that another person makes on an audience (see, for example, Gillespie, 1990; Schlenker & Britt, 1999). No one, to our knowledge, has studied whether or how people try to manage how real they seem to others. Maybe that reality is just taken for granted, because everyone comes conveniently wrapped in a body that can be seen, touched, and so on. As Warriner (1956) and others have noted, however, the body is not what psychologists usually have in mind when they speak of analyzing some person. Psychologists prefer to study inner qualities, such as a person’s thoughts and feelings, which can only be inferred from patterns in that person’s behavior. The same process can be used to detect groups. It seems interesting, though, that the perceived reality of other people has not been studied by psychologists. There certainly are people who seem less real than others, because their behavior is more difficult to understand, or because they have been classified in certain ways (e.g., mentally or physically ill, disabled, retarded, elderly, deviant) that suggest they are “not all there.” Research on ostracism shows that it is painful to be ignored by others, as if one were not real (K.D.Williams & Zadro, 2001), and aggression research shows that people who seem less real to others maybe treated harshly as a result (see Bandura, Underwood, & Fromson, 1975; Gamson, 1995; Kramer, Ber, & Moore, 1987; Silver, Conte, Miceli, & Poggi, 1986). The perceived reality of a person may thus be an important factor in his or her social life, one that merits more consideration. People can also try to manage the impressions that others form about a group. Again, the goal of such behavior is often to make a better personal impression. For example, a person could try to emphasize his or her associations with successful groups, exaggerate the positive qualities of groups with which he or she is associated, distance himself or herself from unsuccessful groups, or exaggerate the negative qualities of groups with which he or she is not associated (see Cialdini, 1989). Efforts to manage impressions about groups are not always self-serving, however. Someone might try to help (or hurt) a group by offering others positive (or negative) information about it, regulating access to such information, or whatever. These are common activities in everyday life (consider corporate spokespersons and public relations firms), but they have been ignored by most social psychologists. The limited work that can be found in this area comes from organizational psychologists instead (see, for example, Benoit, 1997; Dutton, Dukerich, & Harquail, 1994; R.L.Higgins & Snyder, 1989; Mohamed, Gardner, & Paolillo, 1999; Staw, McKechnie, & Puffer, 1983), and that work often focuses on the favorableness of the impressions about groups. Work on other aspects of groups, such as their perceived reality, is even scarcer. Because such work is so rare, a traditional literature review of it was just not feasible. Instead, we simply searched for any theory or research that seemed relevant to managing impressions about the reality of groups. We also discussed the issue with colleagues, who suggested other work that we should consider. In the end, we were pleasantly surprised, because we found more material than expected. We divide that material, in a rather crude way, into two broad categories, namely, work on (a) making groups seem more real than they are and (b) making groups seem less real than they are. MAKING GROUPS SEEM MORE REAL When, how, and why might people try to make groups seem more real than they are? Work on this kind of impression management can be found in several sources.
ENTITATIVITY AND SOCIAL INTEGRATION
315
Problem Identification Every group faces problems occasionally, and some (but not all) of them involve low or declining social integration. Examples of such problems include weak commitment among some members, who may be planning to leave the group; private conflicts between group members; and the impending loss of resources that the group needs in order to succeed. How do groups solve these problems? The identification of a problem is the first, and arguably the most important, phase of group problem solving. Problem identification includes detection and diagnosis—problems are identified when people notice there is something wrong and then develop an explanation for the anomaly. Moreland and Levine (1992a) reviewed theory and research on problem identification by groups and proposed a conceptual model of that process. This model allowed them to develop hypotheses about such issues as when a group problem will be identified and who will identify it. They also identified the options available to someone who identifies a group problem. These options are (a) denying and distorting the problem, (b) hiding the problem from others; (c) waiting and watching the problem to see how it develops; (d) solving the problem independently, without involving other group members; (d) collaborating with other group members to solve the problem; and (f) trying to involve outsiders in some way. Among these six options, (a) and (b) could be interpreted as attempts to make a group seem more real than it is. Option (b), hiding a problem from others, is a clear example in this regard, because if other people (in or out of the group) are not aware that a problem exists, then the group may seem healthier than it is. Put another way, they are likely to overestimate the group’s level of social integration. When would someone try to hide a problem in a group from others? Moreland and Levine (1992a) suggest that this occurs when a person feels responsible for the problem or might be held accountable for it by others. In these cases, guilt or fear might lead the person to hide the problem. Similarly, someone may try to hide a problem when he or she fears that it will weaken morale in the group or create conflict among group members (Janis, 1972). How can a group problem be hidden? Moreland and Levine argue that some one could delay a problem’s detection by preventing symptomatic events from occurring, concealing such events from others, and inducing others to ignore or discount any symptomatic events that they notice. And the diagnosis of a problem could be delayed by offering incorrect explanations for it or by disparaging those who explain the problem correctly. Moreland and Levine claimed that a group problem is often difficult to hide and that it may not stay hidden for long, especially when the problem is familiar to the group or is severe, when the group monitors carefully events both in and around it, or many people have the expertise needed to identify the problem. Option (a), denying or distorting a problem, is a less clear example of managing beliefs about the reality of a group, because the impression manager and his or her audience are the same person. Nevertheless, there may be times when someone is so threatened by a group problem that he or she tries not to acknowledge it. Moreland and Levine suggest that this might happen when a problem seems serious, but also difficult or impossible to solve; when the person feels responsible for causing the problem or might be held accountable for it by others; or when the group is successful, evokes strong commitment in its members, and discourages problem identification. How could someone deny and distort a group problem? Moreland and Levine argue that a person could avoid any situations where symptomatic events are likely to occur, ignore or discount such events when they do occur, and try not to discuss the problem with other group members. The person could also misdiagnose the problem in ways that make it appear less threatening, perhaps by attributing it to external and unstable causes. Other possible tactics include minimizing a problem’s consequences or imagining that it can be solved quickly or easily (see Ross & Staw, 1993; Staw, 1997).
316
THE PSYCHOLOGY OF GROUP PERCEPTION
Group Socialization An important aspect of all groups is that they change over time (see Arrow, McGrath, & Berdahl, 2000). These temporal changes include group formation and dissolution, the development of groups, and group socialization. The latter phenomenon is interesting here, because socialization often tempts people to make groups seem more real than they are. A general model of group socialization, based on a broad review of theory and research in that area, was proposed by Moreland and Levine (1982). According to that model, the relationship between a group and each of its members changes in systematic ways over time. These changes are governed by three basic processes, evaluation, commitment, and role transition. Evaluation involves attempts by the group and the individual to evaluate and maximize one another’s rewardingness. Evaluation produces feelings of commitment, which can rise and fall over time. When commitment reaches a decision criterion, a role transition occurs. The relationship between the group and the individual is trans formed, and both parties begin to evaluate one another again, often in different ways than before. A cycle of socialization activity is thus created, one that moves the individual through the group. According to Moreland and Levine (1982), group membership is thus dynamic, rather than static. People can experience five phases of membership (investigation, socialization, maintenance, resocialization, and remembrance), separated by four role transitions (entry, acceptance, divergence, and exit). Investigation is noteworthy because of the impression management that often occurs in that membership phase. Investigation is a period when a prospective member, who does not yet belong to the group, engages in reconnaissance, evaluating how well the group might satisfy his or her personal needs. At the same time, the group engages in recruitment, evaluating how much that person might contribute to the achievement of its goals. If the group’s evaluations of the individual are positive, then its commitment to that person will rise. At some point, it might cross the group’s decision criterion, and if that happens, then the group will try to produce the role transition of entry, making the person a new member. The more committed a group is to a prospective member (especially when that person is less committed to the group), the more the group will try to persuade the person to join. And that is where the temptation to manage impressions about the group may arise. Any weaknesses in the group, involving low or declining social integration, may be hidden, and if such problems are discovered, then they may be discounted. The group may also try to provide the prospective member with “evidence” that its level of social integration is high. For example, the person might be told or shown that group members spend a lot of time together, or experience both process and outcome interdependence in their daily activities. Without these and other tactics, there is a risk that the prospective member will not want to enter the group; he or she might even join another group instead (see Levine, Moreland, & Ryan, 1997). Many researchers have found that groups indeed try to manage the impressions they make on prospective members (see Wanous, 1989; Zander, 1976), and one goal of such efforts (but not the only one) is to make groups seem real. When prospective members enter such groups and become new members, they may thus have overly optimistic expectations about what their groups will be like (see Brinthaupt, Moreland, & Levine, 1991; Pavelchak, Moreland, & Levine, 1986; Wanous, 1976; see also Ryan & Bogart, 1997). Expectations like these are dangerous, because as newcomers discover the reality of group life, they are likely to be disappointed, and they may be ill prepared to cope with whatever problems the group presents to them. This may lead some newcomers to leave their groups—turnover rates do tend to be higher during the first few weeks of group membership (see Fichman & Levinthal, 1991; Luke, Roberts, & Rappaport, 1993; Wanous, 1980). One tactic a group might use to avoid this problem is to be more frank with prospective members, providing them with realistic “previews” of what group membership is really like (see Kaul & Bednar, 1994; Phillips, 1998). A preview may drive away some prospective members who do not choose to
ENTITATIVITY AND SOCIAL INTEGRATION
317
enter a troubled group, but those may be the same individuals who would have left the group anyway after they joined and discovered its problems. Dissolving Groups For a variety of reasons, a group can become difficult to maintain, and at some point it may start to dissolve. When that happens, both the people who belong to that group, and the social network in which the group is embedded, may react strongly, often negatively. Efforts may thus be made to conceal the group’s dissolution by making it seem healthier than it is. Evidence for such efforts can be found in several areas of research. Consider, for example, research on divorce and its effects. Marital problems often occur in families containing children, and in such cases parents may try to hide the problems from their children, because they fear the psychological damage that can afflict children whose parents are in conflict (see J.B.Kelly, 2000; Seltzer, 1994; D.S.Shaw, 1992). For example, parents may try not to fight in front of their children, or they may show for each other more affection than they really feel when their children are around. A related area of research involves married or unmarried couples whose relationships are dissolving. In such relationships, one partner may try to hide his or her dissatisfaction from the other. Some evidence indicates, for example, that men are surprised more often than women when relationships dissolve (Hill, Rubin, & Peplau, 1976; Kayser, 1993). This is often attributed to insensitivity among men; the signs of a breakup were there, but they were unnoticed. However, it is also possible than women are more likely than men to hide their dissatisfaction with a relationship, because they are more frightened of confrontation and want to postpone it as long as possible, or because they need more time to prepare for life without their partner. Or men may be less willing than women to acknowledge that a relationship is dissolving. At some point, of course, both partners must realize that their relationship is over. When that happens, they sometimes try to conceal the problem from relatives or friends, because they do not want anyone with a stake in their relationship to feel disappointed or sad, or because they fear that others may try to intervene and thus make the situation even more difficult (Goldsmith & Parks, 1990; Surra, 1988; Parks & Eggert, 1991). What could be done to conceal a dissolving relationship? A couple might try to restrict open conflict (e.g., arguments, criticisms, and sarcastic remarks) to private venues. In public the couple might display instead behaviors analogous to the reflection and labeling that we mentioned earlier. Reflection, or behaving like a couple, might include going places together, showing mutual affection, making joint decisions, or agreeing about important issues. Labeling, or talking like a couple, might include using collective rather than personal pronouns (us, we, and our, not I, me, or my—see Sillars, Shellen, McIntosh, & Pomegranate, 1997; Veroff, Sutherland, Chadiha, & Ortega, 1993). An interesting issue is how long a dissolving relationship can be concealed. It must be difficult for partners to pretend that nothing is wrong, especially when their relationship is weak. And people who know the couple well, interact with them often, or have a stake in their relationship may be highly sensitive to signs of trouble. Yet another area of relevant research involves work groups or organizations that are dissolving, usually because they are unprofitable. The leaders of such groups may try to conceal their decisions to end them, because there is still work to be done. Leaders fear that once workers learn the truth, they will quickly leave the group, or at least lose some commitment to it, behave in a variety of harmful ways (e.g., arrive at work late or skip it altogether, steal money or goods, damage valuable equipment, reveal sensitive information to outsiders), and so on (see Greenhalgh, 1983). Efforts to conceal the dissolution of work groups might include limiting access by workers to information about group performance; attributing poor performance to external or unstable factors, or acting as if poor performance is unimportant; refusing to answer workers’
318
THE PSYCHOLOGY OF GROUP PERCEPTION
questions about the future of their groups, or answering such questions deceptively; controlling rumors that groups are dissolving; and offering misleading explanations for any events (e.g., layoffs, transfers of people or goods, sales of equipment) that could be interpreted by workers as signs that their groups might be dissolving. Ironically, Sutton (1987) and others have found that such efforts may be unnecessary. Many workers react positively to the news that their groups might soon dissolve, working harder and longer than usual, perhaps because they believe that those groups can still be saved, or that their work will be noted and they will be transferred to other groups (or receive positive recommendations that can help them to obtain new jobs). And Sutton argues that there may be ways to manage the dissolution of a work group that will minimize the confusion and pain its members may feel. For example, helping workers complete the tasks of disbanding (ending relationships based on work group membership) and reconnecting (forming new relationships with the same people that do not revolve around work) can be especially valuable. Political Behavior Finally, the political behavior of some groups (and their members) provides further examples of efforts to make groups seem more real. In several areas of the world, for example, large groups of people (e.g., the Palestinians, the Kurds, and the Basque) desperately want to become nations. An important step toward that goal is to increase the entitativity of those groups so that they are viewed as nationalities by others. Klein, Azzi, Brito, & Berckmans (1999) describe several tactics that can be useful in this regard. These include emphasizing that the people have (a) a common ancestry; (b) a shared ideology; (c) outcome interdependence; (d) a territory; (e) a distinct language; (f) a similar psychological makeup; and (g) strong bonds. Becoming a nation can be rewarding in many ways. Aside from greater political freedom, the citizens of new nations may enjoy both material (e.g., economic opportunities) and psychological (e.g., pride, security) gains. Similar gains may be experienced by people living within an existing nation who persuade others that they have legitimate needs that are not yet met. Schneider and Ingram (1993) argue that such people must first be viewed as a group; only then will their rights be adequately identified and assessed. A similar example involves small groups that are embedded within larger organizations. Given that all organizations have limited resources, groups within those organizations must compete with one another for whatever resources are available. Such competitions are more likely to be won by groups that seem healthier to those who make decisions about resource allocations. Thus, the most successful groups should be those that can manage their entitativity effectively. Evidence for that claim can be found in the work of Ancona and Caldwell (1988; 1992), who believe that success for most work groups depends more on their control of external than internal factors. In studies of new product development teams, they often noted special roles, whose apparent purpose was to control the flow of resources (e.g., people, materials, funds, information) into and out of those groups. One of those roles, the ambassador, involved efforts by some group member to go out into the organization and “sell” the group so that it might receive more resources. Indeed, groups with ambassadors proved to be more successful than those without them. One way to sell a group, of course, is to make it seem healthy, whatever its real health may be. This might be accomplished by claiming that a group is cohesive and that its members have similar values (for example). Political groups often compete with each other for members (Levine, Moreland, & Ryan, 1997), and for other resources, such as financial support, in the communities where they operate (see McPherson, 1983; Simon & Klandermans, 2001). One tactic that the leaders of such groups might use is to make their groups seem healthier than they are (see Pfeffer, 1981). A leader might, for example, exaggerate the size of a group, the commitment of its members, or the group’s accomplishments (e.g., taking credit for events or outcomes
ENTITATIVITY AND SOCIAL INTEGRATION
319
for which the group was only partly responsible; see Oots, 1989). These tactics might well impress the community, making it more likely to support the group.4 And group members, insofar as they are unsure about their group’s actual level of social integration, might be encouraged, so they would be less likely to quit the group, feel more committed to it, and work harder on its behalf, helping it to accomplish more. Finally, the members of other, competing groups may find such messages discouraging. This leads to a related issue, namely what leaders might say about the health of other groups with which their own groups are competing. Leaders sometimes choose to make competing groups seem healthier than they actually are. The main value of this tactic is to alarm members of the leaders’ own groups (see Abelson, Dasgupta, Park, & Banaji, 1998; Coovert & Reeder, 1990; Insko & Schopler, 1998), thereby strengthening the cohesion of those groups, inhibiting deviance among group members, and justifying more authoritarian behavior by the leaders (see Middents, 1990; Staw, Sandelands, & Dutton, 1981; Sherif, Harvey, White, Hood, & Sherif, 1961). But a more common tactic among leaders may be to make rival groups seem less healthy than they are. That would weaken support in the community for those groups, discourage their members (insofar as they are unsure about the actual levels of social integration in their own groups), and even make members of the leaders’ own groups more optimistic about their chances for competing successfully. MAKING GROUPS SEEM LESS REAL Sometimes people try to make groups seem less real than they are. When, why, and how does that type of impression management occur? Once again, relevant work can be found in several sources. Extramarital Affairs Love affairs between married and single persons, or between people who are married to different persons, are disturbingly common. Although the true frequency of such affairs is unknown, because researchers only know about the relationships that people are willing to describe, Richardson (1986) has estimated that nearly half of all married men are unfaithful to their wives at some point during their marriages. Men often try to conceal such affairs, because they fear the painful consequences of exposure. Those consequences include a weakening of marital bonds, sometimes leading to (a costly) divorce; psychological harm to wives and children, and public shame. How are extramarital affairs hidden? Richardson found that the most common tactics were cloaking and compartmentalization. Cloaking is accomplished by scheduling normal activities, especially work activities, in ways that allow the participants in an extra-marital affair to spend lots of time together without arousing too much suspicion. This tactic is interesting, because rather than concealing the relationship altogether, it makes one kind of relationship (love) appear to be another relationship (work) whose level of social integration is much lower. Compartmentalization is accomplished by separating someone’s love life from his or her married life. For example, participants in an extramarital affair might limit their love life to settings (e.g., restaurants, bars, and hotels that are small or unpopular, other cities) where spouses, relatives, and friends are unlikely to witness their infidelity. Office Romances Love affairs sometimes occur within work settings (Powell & Foley, 1999). This has created a variety of moral and legal problems for business organizations, which typically have no formal policies about such affairs (L.Hoffman, Clinebell, & Kilpatrick, 1997). There is general agreement, however, that office
320
THE PSYCHOLOGY OF GROUP PERCEPTION
romances can cause many problems, such as lost or wasted time, strained relationships with coworkers, and even exploitation, when older, higher status men become involved with younger, lower status women (see G.E.Jones, 1999). As a result, office romances are discouraged at most workplaces, though seldom forbidden. Workers involved in these relationships are thus motivated to hide them from managers and coworkers (C.I.Anderson & Hunsaker, 1985; Quinn, 1977). Several tactics can be used to achieve that goal, but most researchers have studied how office romances are discovered, rather than concealed. Dillard and Miller (1988) found, for example, that the clues used to discover office romances include being seen together often outside of work, spending an unusual amount of time together at work, and taking very long lunch breaks together. Similar clues were found by Anderson and Hunsaker (1985) and Quinn (1997), but they identified many others as well, such as unusual happiness at work, favoritism in task assignments and evaluations, and poor task performance. These clues provide an indirect means of identifying the tactics that participants in office romances could use to conceal their relationships. While away from work, for example, they might try not to be seen together; while at work, they might try to make unbiased, or even negatively biased, task assignments and evaluations. Some of the researchers in this area, such as Anderson and Hunsaker (1985), Hoffman, Clinebell, and Kilpatrick (1997), and Quinn (1977) claim that office romances are difficult or impossible to conceal. Participants often believe that their relationships are secret, but managers and coworkers usually realize what is happening (and pretend to be unaware of it). This is a difficult claim to test, however, because researchers only know about office romances that people are willing and able to describe to them. Nothing is known about office romances that went undetected, or about the tactics participants used to conceal those relationships. Gay and Lesbian Relationships Although acceptance of homosexuality is greater now than ever before, gay and lesbian people must still endure hardships, including limited job opportunities and security, strained relations with their families, and social stigma. So it is not surprising that the participants in such relationships often try to conceal them from others (Kitzinger, 1991). Researchers have identified many tactics that people can use to achieve that goal. For example, Kitzinger (1991) found that homosexuality may be concealed by (a) dressing and acting in ways stereotypical of one’s own sex; (b) making appreciative comments about a member of the opposite sex; (c) claiming a boyfriend (a lesbian tactic) or girlfriend (a gay tactic) who lives far away and thus is seldom seen; (d) arranging for a “beard”—someone of the opposite sex (often gay or lesbian as well) who poses as a boyfriend or girlfriend on “dates,” at parties, and so on; and (e) introducing actual lovers as “friends.” Casual conversations about life at home should also be avoided, because it is easy to slip up and provide other people with clues about one’s sexual orientation. In fact, gays and lesbians often avoid close relationships with straight people, because such relationships increase the risk that their true sexual orientations will be discovered. Davies (1992) found that homosexuality can be concealed using compartmentalization, relocation, and collusion. Compartmentalization involves keeping gay or lesbian activities completely separate from other aspects of one’s life. Davies found someone, for example, whose family lived in one town, who lived and worked in another town, and who confined his sexual activities to yet a third town. This person also exemplifies relocation—moving away from the people someone cares about most so that they never see any “improper” behavior. Collusion involves revealing homosexuality to a close friend, who can then help someone conceal his or her sexual orientation from others. Davis described an interesting example of this
ENTITATIVITY AND SOCIAL INTEGRATION
321
tactic as well—a person whose parents knew all about his homosexuality, but helped him to conceal it from his grandparents. Finally, Ponse (1976) identified many of the same tactics as Kitzinger (1991), but she found several others as well. For example, she noted that homosexuals must often resist reacting to detrimental remarks by others about gays or lesbians. And they must be very careful about reacting to come-ons that they may receive from other homosexuals when heterosexuals are present. Ponse also analyzed what she called “counterfeit secrecy” in the lives of some homosexuals, who collude with their families and friends to pretend that they are heterosexual, even though everyone knows that is untrue. This tactic is a form of denial, one that can become a burden as time goes on and further evidence of someone’s homosexuality accumulates but must be ignored. Subversive Groups Work on extramarital affairs, office romances, and gay or lesbian relationships involves dyads rather than larger groups. And efforts to manage beliefs about the reality of those groups are aimed at denying their existence altogether rather than making them seem less real. A different area of research, closer in spirit to the work we reviewed earlier on making groups seem more real, involves various subversive groups, such as gangs, cults, partisans, or terrorists. These groups often want to avoid public attention, because that can bring pressure from authorities, who may try to weaken or dissolve them. How can subversive groups be made to seem less real than they are? Research on such groups is difficult, for obvious reasons, but some information is available (Blee, 2002; Crenshaw, 1985; della Porta, 1995; Lyman & Scott, 1967; McCauley & Segal, 1987; Oots, 1989). A common tactic for concealing a subversive group is to “go underground” by dividing the group into smaller cells, each of which has little contact with the others. Cell meetings are held secretly, and information about the group is distributed on a “need to know” basis. An image of “leaderless resistance” is promoted, so that the group’s structure becomes unclear, especially to outsiders. Some groups may even attempt to persuade authorities that they have already dissolved, so that there is no longer a need to be concerned about them. As della Porta (1995) has noted, this tactic is dangerous, because if authorities believe that the group is weak, or even dissolved, then the community where that group operates may come to the same conclusion. Many subversive groups rely on their communities for important resources, such as people, materials, funds, and information. So a subversive group that pretends to be weak might actually weaken itself through a kind of self-fulfilling prophecy. The real trick for such groups is to avoid attracting the attention of authorities while not losing a community’s interest and potential support. WHO IS FOOLING WHOM? Clearly, there are many tactics that people can use to manage beliefs about the reality of groups. But another aspect of the work we reviewed was also intriguing, namely, the wide variety of sources and targets that can be involved in such impression management. A summary of those sources is provided in Figure 22.2. The rows in that figure are possible sources, while the columns are possible targets. The first five rows (A to E) are all ingroup sources. Three of these are individuals—oneself (A), one’s partner in a dyadic relationship (B), and the groups leader (C). The other two are groups—a clique within the group (D) and the ingroup as a whole (E). The next three rows (F to H) are all outgroup sources. One of these is an individual—the outgroup’s leader (F). The other two are groups—a clique within the outgroup (G) and the outgroup as a whole (H). Finally, the last three rows are sources in the group’s environment.
322
THE PSYCHOLOGY OF GROUP PERCEPTION
Source/Target
A
B
C
D
E
F
G
H
I
J
K
A. Self B. Partner C. Leader D. Clique E. Ingroup F. Leader G. Clique H.Outgroup I. Stakeholders J. Authorities K. Community FIGURE 22.2. Sources and targets of impression management about groups
Many groups, for example, are embedded in organizations of various kinds. Setting sources include individual or group stakeholders (I) and authorities (J), and the community where the group operates (K). Stakeholders are people (e.g., friends or relatives of group members, prospective or ex-members of the group, or investors) with some interest in the group’s outcomes. Authorities (e.g., administrators, managers, politicians, police) are people who monitor or regulate the group or allocate resources to it. These descriptions of the rows in Figure 22.1 apply to its columns as well, of course. We have already provided examples for many of the cells in the figure, but not for all of them. The advantage of such a figure is that it draws attention to situations that have not been studied, such as efforts by one leader to manage the beliefs of another leader about a group, or efforts by one clique within a group to manage the beliefs of another clique in the same group. One interesting issue that the figure raises is whether people are managing impressions about their own group or about another group to which they do not belong. The latter group may be an outgroup, with which the ingroup often competes, or simply a group from the same environment as the ingroup. Another interesting issue raised by the figure is whether the source or the target of impression management is a person or a group (cf. Latané, 1981). When the source is a group (e.g., a married couple, a clique within a group, or the group as a whole), coordination problems may arise. Do people plan ahead in such situations, so that they can collaborate more effectively in their efforts to manage the beliefs of others about a group? If not, then what would happen if one person said or did something that undercut the impressions others were trying to make (cf. Wood, Lundgren, Ouellette, Busceme, & Blackstone, 1994)? When the target is a group, how are efforts at impression management distributed across targets? Are some people in the target group more important than others because of their status? If so, then more effort ought to be directed toward managing their beliefs, but how would other members of the group react to that tactic? These are issues that deserve more consideration. FINAL REMARKS We began this chapter with a look back at Campbell’s (1958) paper on entitativity, a concept that has recently become popular among social psychologists who study groups. We argued that entitativity, as Campbell envisioned it, refers not only to whether groups seem real—but also whether groups are real. To avoid any more confusion about this distinction, we suggested that social integration (see Moreland, 1987)
ENTITATIVITY AND SOCIAL INTEGRATION
323
be used as a term for the latter meaning of entitativity. Social integration and entitativity are correlated positively in most cases, we believe, because groups that are more real tend to seem more real as well. But the correlation is not perfect, so groups can seem more (or less) real than they are. As a result, impression management about the reality of groups is possible. We reviewed whatever work we could find on how, when, and why people attempt to manage beliefs about the reality of groups. There was more such work than we expected, and it revealed many different impression management tactics. These often reflected the factors that Campbell (1958) identified as likely sources of entitativity, such as proximity, similarity, and common fate. But many of the tactics that we found reflected other factors as well. Our review raised several broader issues about the general process of managing beliefs about the reality of groups. Three of these seem especially interesting. First, any effort to manage beliefs about a group obviously involves some source and target (as noted earlier) and a group. Regarding the source, some people are probably more skilled than others at this kind of impression management. Who? Maybe the people who are effective at managing impressions about individuals (see Schlenker & Weigold, 1992) are also good at managing impressions about groups. And maybe it is helpful if the source actually believes what he, she, or it is trying to persuade the target about. Regarding the target, some people are probably more vulnerable than others to this kind of impression management. Who? Familiarity or contact with the group, or membership in similar groups, is probably important (see Brewer & Harasty, 1996; Gaertner & Schopler, 1998; M.P.Johnson & Milardo, 1984; McGarty, Haslam, Hutchison, & Grace, 1995; Ponse, 1976). For example, Johnson and Milardo, and Ponse, argued that belonging to one group can help someone to see other groups of the same kind. Personality factors, such as the need for structure (see Brewer & Hardasty, 1996; Moskowitz, 1993; Neuberg & Newson, 1993) may be important too; organizing the world into groups, rather than individuals, can simplify social perception and make things seem more predictable. Moods can be a factor as well; happy people are likely to categorize others more broadly (see Dovidio, Isen, Guerra, Gaertner, & Rust, 1998; Queller, Mackie, & Stroessner, 1996). And people from Eastern cultures, for whom groups may be more important (see Markus & Kitayama, 1991), may be better than others at gauging how real a group is and thus be less affected by efforts to manage their beliefs about that group. Finally, regarding the group itself, some groups may naturally seem more real than others because of their composition (see Brewer & Harasty, 1996; Hamilton, Sherman, & Lickel, 1998; McGarty, et al., 1995; Moreland & Levine, 1992b), structure (see Hamilton, et al., 1998; McConnell, Sherman, & Hamilton, 1997), or activities (see Brewer, 1979; Gaertner & Schopler, 1998; Lindskold, Albert, Baer, & Moore, 1976; Weick & Roberts, 1993). Beliefs about the reality of such groups should thus be more difficult to manage. Another broad issue that interests us is the relative difficulty of making groups appear more, versus less, real than they are. Is it easier, for example, to persuade people that a group exists when it does not, or to persuade them that a group does not exist when it does? In our opinion, entitativity is more easily raised than lowered. Why? One reason is that human beings appear to be group creatures by nature (see Baumeister & Leary, 1995; Caporael & Baron, 1997; Insko, Schopler, & Sedikides, 1998). As a result, they may be alert to the presence of groups, because of their evolutionary value. From an evolutionary perspective, it might be better to “see” groups that are not there, rather than to overlook groups that are there. Another reason for our opinion is work indicating that people (a) often “see” more groups in a setting than are really there and (b) seldom overlook groups that are there. Regarding (a), research by Freeman (1992) and others on how well people can recall social structure information shows that intrusion errors are common—people recall more groups than were actually there. And consider the apparent ease with which minimal groups researchers can convince people that they belong to (novel, meaningless) groups, and that
324
THE PSYCHOLOGY OF GROUP PERCEPTION
membership in such groups is a potential source of pride (see Sherman, Hamilton, & Lewis, 1999). As for (b), research on efforts by people to hide extramarital affairs, office romances, and gay or lesbian relationships suggests that such efforts often fail (see Anderson & Hunsaker, 1985; Hoffman, Clinebell, & Kilpatrick, 1997; and Quinn, 1977). Even subtle signals, such as using a soft tone of voice, standing somewhat closer to someone, or monitoring carefully how others behave toward a person (or what they say about him or her), are enough in many cases for a concealed relationship to be discovered. Finally, time is always an important factor in social behavior, and so its potential role in managing beliefs about the reality of groups ought to be considered. Campbell (1958), for example, argued that a groups entitativity can change over time (perhaps due to changes in the group’s level of social integration). That raises the issue of how prior beliefs about the reality of a group might affect current beliefs about its reality. How do people react to increases or decreases in entitativity in their own or other groups? If groups are indeed so important to people as we suggested, then it seems likely that people would react more strongly to changes in the entitativity of their own groups than to similar changes in other groups. And in their own groups, decreases in entitativity should evoke stronger reactions than increases in entitativity, whereas the opposite pattern should occur for other groups. What about groups that have dissolved and thus are no longer “real” in the usual sense of that term? Common sense suggests that groups can continue to affect people long after they have dissolved, and our own research on loyalty (Moreland & McMinn, 1999) shows that even minimal groups have such power. What about groups that have not yet formed, and thus are not “real” either? Some people, such as visionary lead ers, matchmakers, or protective parents, seem able to “see” potential groups (see Parks, Stan, & Eggert, 1983). In some cases, they can even predict when and how those groups might form.5 That lets them, if they want to, meddle in ways that either help such groups to form or hinder their formation. There is much excitement now about group entitativity. We hope that this chapter will contribute to that excitement. But the chapter has another purpose as well, at least in our minds. Much of the theory and research on entitativity thus far reflects a social cognition perspective. That is not surprising when entitativity is taken to mean how real a group seems to people. But as we have noted, there is more to a group than mere appearance. In fact, appearance and reality can diverge, especially when efforts to manage beliefs about groups are made. It would be disappointing, in our opinion, if work on entitativity goes the same way that minority influence work has gone (Levine & Kaarbo, 2001). In minority influence research, social interaction is seldom studied now, so many intriguing phenomena involving the actual struggle between minorities and majorities are overlooked or ignored. We hope our chapter will serve as a gentle reminder that entitativity has a social aspect that ought to be studied. NOTES 1. An example of that debate, as it applies to the “chemistry” that arises among the members of some work groups, can be found in Moreland, Levine, & Wingert (1996). 2. We wonder whether this explains the unusual quote by Ciardi that editors of the journal chose to fill a bit of empty page space at the end of Campbell’s paper. 3. In fact, recent research in our laboratory suggests that people view cliques as strong evidence that a “group” is not really a group at all. 4. Several years ago, in a research project on campus groups, one of us discovered that several groups of a similar kind actually had many of the same members. Apparently, this was an effort by those groups to acquire more funds overall from student government, which always counted the number of students who belonged to each campus group, but rarely compared lists with one another.
ENTITATIVITY AND SOCIAL INTEGRATION
325
5. The fact that dissolved and unformed groups are both on peoples minds could be viewed as further evidence for the importance of groups to our species.
References
Aarts, H., & Dijksterhuis, A. (2001). The silence of the library: Environment, situational, nor, and social behavior. Manuscript submitted for publication. Abelson, R.P. (1959). Modes of resolution of belief dilemmas. Journal of Conflict Resolution, 3, 343–352. Abelson, R.P., Dasgupta, N., Park, J., & Banaji, M.R. (1998). Perceptions of the collective other. Personality and Social Psychology Review, 2, 243–250. Abelson, R.P., & Rosenberg, M.J. (1958). Symbolic psycho-logic: A model of attitudinal cognition. Behavioral Science, 3, 1–13. Abrams, D. (1985). Focus of attention in minimal intergroup discrimination. British Journal of Social Psychology, 24, 65–74. Abrams, D. (1990). How do group members regulate their behaviour? An integration of social identity and selfawareness theories. In D.Abrams & M.A.Hogg (Eds.), Social identity theory: Constructive and critical advances (pp. 89–112). London and New York: Harvester Wheatsheaf and Springer Verlag. Abrams, D. (1992). Processes of self-identification. In G.Breakwell (Ed.), Social psychology of identity and the selfconcept (pp. 57–99). San Diego: Academic Press. Abrams, D. (1994). Social self-regulation. Personality and Social Psychology Bulletin, 20, 473–483. Abrams, D. (1996). Social identity, self as structure and self as process. In W.P.Robinson (Ed.), Social groups and identities: Developing the legacy of Henri Tajfel (pp. 143– 167). Oxford: Butterworth-Heinemann. Abrams, D. (1999a). Social identity, social cognition, and the self: The flexibility and stability of self-categorization. In D.Abrams & M.A.Hogg (Eds.), Social identity and social cognition (pp. 197–209). Maiden, MA: Blackwell. Abrams, D. (1999b). Social and psychological dynamics within and between groups: How far do reality and perception coincide? Society for Experimental Social Psychology, Preconference on Group and Intergroup Processes. St. Louis, October. Abrams, D., & Brown, R.J. (1989). Self-consciousness and social identity: Self-regulation as a group member. Social Psychology Quarterly, 52, 311–318. Abrams, D., & Hogg, M.A. (1988). Comments on the motivational status of self-esteem in social identity and intergroup discrimination. European Journal of Social Psychology, 18, 317–334. Abrams, D., & Hogg, M.A. (1990). An introduction to the social identity approach. In D.Abrams & M.A.Hogg (Eds.) Social identity theory: Constructive and critical advances (pp. 1–9). London and New York: Harvester Wheatsheaf and Springer-Verlag. Abrams, D., & Hogg, M.A. (1998). Prospects for research in group processes and intergroup relations. Group Processes and Intergroup Relations, 1, 7–20. Abrams, D., & Hogg, M.A. (Eds.) (1999). Social identity and social cognition. Oxford: Blackwell. Abrams, D., & Hogg, M.A. (2001). Collective identity: Group membership and self-conception. In M.A.Hogg & R.S.Tindale (Eds.), Blackwell handbook of social psychology: Group processes (pp. 425–460). Oxford: Blackwell. Abrams, D., Marques, J.M., Bown, N., & Dougill, M. (2002). Anti-norm and pro-norm deviance in the bank and on the campus: Two experiments on subjective group dynamics. Group Processes and Inter group Relations, 5, 163–182. Abrams, D., Marques, J.M., Bown, N., & Henson, M. (2000). Pro-norm and anti-norm deviance within and between groups. Journal of Personality and Social Psychology, 78, 906–912. Abrams, D., & Masser, B. (1998). Context and the social self-regulation of stereotyping: Perception, judgment, and behavior. In R. S.Wyer (Ed.), Advances in social cognition (Vol. 11, pp. 53–68). Hillsdale, NJ: Erlbaum. Adorno, T.W., Frenkel-Brunswik, E., Levinson, D.J., & Sanford, R.N. (1950). The authoritarian personality. New York: Harper.
REFERENCES
327
Alexander, M.G., Brewer, M.B., & Herrmann, R. (1999). Images and affect: A functional analysis of out-group stereotypes. Journal of Personality and Social Psychology, 77, 78– 93. Alexander, R.D. (1974). The evolution of the social behavior. Annual Review of Ecology and Systematics, 5, 324–382. Allison, S.T., & Messick, D.M. (1985). The group attribution error. Journal of Experimental Social Psychology, 21, 563–579. Allport, F.H. (1924). Social psychology. Boston, MA: Houghton Mifflin. Allport, F.H. (1962). The structoronomic concept of behavior: Individual and collective: 1. Structural theory and the master problem of social psychology. Journal of Abnormal and Social Psycholoyg, 64, 1–30. Allport, G. (1954). The nature of prejudice. Reading, MA: Addison-Wesley. Allport, G.W., & Postman, L. (1947). The psychology of rumor. New York: Henry Holt & Company. Altemeyer, B. (1988). Enemies of freedom: Understanding right-wing authoritarianism. San Francisco: Jossey-Bass. Alwin, D.F., Cohen, R.L., & Newcomb, T. M. (1991). Political attitudes over the life span: The Bennington women after fifty years. Madison: University of Wisconsin Press. Ancona, D.G., & Caldwell, D.F. (1988). Beyond task and maintenance: Defining external functions in groups. Group and Organization Studies, 13, 468–494. Ancona, D.G., & Caldwell, D.F. (1992). Bridging the boundary: External activity and performance in organizational teams. Administrative Science Quarterly, 37, 634–665. Andersen, C.A. (1997). Effects of violent movies and trait irritability on hostile feelings and aggressive thoughts. Aggressive Behavior, 23, 161–178. Andersen, S.M., & Ross, L. (1984). Self-knowledge and social inference: I. The impact of cognitive/affective and behavioral data. Journal of Personality and Social Psychology, 46, 280–293. Anderson, B. (1991). Imagined communities. Reflections on the origin and spread of nationalism. London: Verso. Anderson, C.I., & Hunsaker, P.L. (1985, February). Why there’s romancing in the office and why it’s everybody’s problem. Personnel, 59–63. Anderson, J.R. (2000). Cognitive psychology and its implications. New York: Worth Publishers. Anzulovic, B. (1999). Heavenly Serbia: From myth to genocide. London: Hurst & Company. Arend, S. (2000). The effects of perceived group meaningfulness on ingroup bias. Unpublished honors thesis, University of Queensland, Australia. Aron, A., Aron, E.N., & Smollan, D. (1992). Inclusion of other in the self scale and the structure of interpersonal closeness. Journal of Personality and Social Psychology, 63, 596–612. Aronson, E., & Mills, J. (1959). The effect of severity of initiation on liking for a group. Journal of Abnormal and Social Psychology, 59, 177–181. Arrow, H., McGrath, J.E., & Berdahl, J.L. (2000). Small groups as complex systems: Formation, coordination, development, adaptation. Thousand Oaks, CA: Sage. Asch, S.E. (1940). Studies in the principles of judgments and attitudes: II. Determination of judgments by group and by ego standards. Journal of Social Psychology, 12, 433–465. Asch, S.E. (1952). Social psychology. Englewoof Cliffs, NJ: Prentice-Hall. Asch, S.E. (1956a). Opinions and social pressure. Scientific American, 193, 31–35. Asch, S.E. (1956b). Studies of independence and conformity: A minority of one against a unanimous majority. Psychological Monographs, 70, pp. 1–70. Ashmore, R.D., & Del Boca, F.K. (1981). Conceptual approaches to stereotypes and stereotyping. In D.L.Hamilton (Ed.), Cognitive processes in stereotyping and intergroup behavior (pp. 1–35). Hillsdale, NJ: Erlbaum. Askevis-Leherpeux, F., & Bastounis, M. (1998). Differences in ingroup-outgroup representations: An empirical examination of the familiarity hypothesis. Revue Internationale de Psychologie Sociale, 11, 19–31. Atran, S. (1990). Cognitive foundations of natural history. Cambridge: Cambridge University Press and Maison des Sciences de l’Homme. Attneave, F. (1957). Transfer of experience with a class-schema to identification-learning of patterns and shapes. Journal of Experimental Psychology, 54, 81–88. Attneave, F. (1959). Applications of information theory to psychology. New York: Holt, Rinehart, & Winston.
328
REFERENCES
Axelrod, R. (1984). The evolution of cooperation. New York: Basic Books. Badea, C., & Askevis-Leherpeux, F. (2002). Perceived group variability and social context: The effect of ingroup size and intergroup comparison dimension. Poster presented at the 13th General Meeting of the European Association of Experimental Social Psychology, San Sebastian, Spain. Bales, R.F. (1953). The equilibrium problem in small groups. In T.Parsons, R.Bales & E.Shils (Eds.), Working papers in the theory of action. New York: Free Press. Banaji, M.R., & Prentice, D.J. (1994). The self in social contexts. Annual Review of Psychology, 45, 297–332. Bandura, A., Underwood, B., & Fromson, M. E. (1975). Disinhibition of aggression through diffusion of responsibility and dehumanization of victims. Journal of Research in Personality, 9, 253–269. Barber, B.R. (1995). Jihad vs. McWorld. New York: Random House. Barker, M. (1981). The new racism. London: Junction Books. Baron, R.M., & Misovich, S.J. (1999). On the relationship between social and cognitive modes of organization. In S.Chaiken & Y. Trope (Eds.), Dual-process theories in social psychology. New York: Guilford Press. Bar-Tal, Y., & Guinote, A. (2002). Who exhibits more stereotypical thinking? The effect of need and ability to achieve cognitive structure on stereotyping. European Journal of Personality, 16, 313–331. Bartlett, F.C. (1932). Remembering. Cambridge: Cambridge University Press. Bartsch, R.A., & Judd., C.M. (1993). Majority-minority status and perceived ingroup variability revisited. European Journal of Social Psychology, 23, 471–483. Bashevkin, S.B. (1991). True patriot love: The politics of Canadian nationalism. Toronto: Oxford University Press. Bassok, M., & Trope, Y. (1984). People’s strategies for testing hypotheses about another’s personality: Confirmatory or diagnostic? Social Cognition, 2, 199–216. Batchelder, W H., & Riefer, D.M. (1999). Theoretical and empirical review of multinomial processing tree modeling. Psychonomic Bulletin and Review, 6, 57–86. Baumeister, R.F. (1982). A self-presentational view of social phenomena. Psychological Bulletin, 91, 3–26. Baumeister, R.F. (1987). How the self became a problem: A psychological review of historical research. Journal of Personality and Social Psychology, 52, 163–176. Baumeister, R.F. (1999). The self in social psychology. Philadelphia: Psychology Press. Baumeister, R.F., & Leary, M.R. (1995). The need to belong: Desire for interpersonal attachments as a fundamental human motivation. Psychological Bulletin, 117, 497– 529. Beauvais, C., & Spence, J.T. (1987). Gender, prejudice, and categorization. Sex Roles, 16, 89–100. Becker. E. (1962). The birth and death of meaning. New York: Free Press. Becker. E. (1973). The denial of death. New York: Free Press. Becker. E. (1975). Escape from evil. New York: Free Press. Benoit, P.J. (1997). Telling the success story: Acclaiming and disclaiming discourse. Albany, NY: SUNY Press. Berger, C.R. (1987). Communicating under uncertainty. In M.E.Roloff & G.R.Miller (Eds.), Interpersonal processes: New directions in communication research (pp. 39– 62). Newbury Park, CA: Sage. Berger, C.R., & Bradac, J.J. (1982). Language and social knowledge: Uncertainty in interpersonal relations. London: Edward Arnold. Berger, P.L., & Luckmann, T. (1966/1971). The social construction of reality. Harmondsworth, UK: Penguin. Berkowitz, L., & Howard, R. (1959). Reaction to opinion deviates as affected by affiliation need and group member interdependence. Sociometry, 22, 81–91. Berman, J.S., & Kenny, D.A. (1976). Correlational bias in observer ratings. Journal of Personality and Social Psychology, 34, 263– 273. Berndsen, M., McGarty, C., van der Pligt, J., & Spears, R. (2001). Meaning-seeking in the illusory correlation paradigm: The active role of participants in the categorization process. British Journal of Social Psychology, 40, 209–233. Berndsen, M., & Spears, R. (1997). Reinterpreting illusory correlation: From biased covariation to meaningful categorisation. Swiss Journal of Psychology, 57, 127–138. Berndsen, M., Spears R., McGarty C., & van der Pligt, J. (1998). Dynamics of differentiation: Similarity as the precursor and product of stereotype formation. Journal of Personality and Social Psychology, 74, 1451– 1463.
REFERENCES
329
Berndsen, M., Spears, R., van der Pligt, J., & McGarty, C. (1999). Determinants of intergroup differentiation on the illusory correlation task. British Journal of Psychology, 90, 201–220. Berndsen, M., Spears, R., & van der Pligt, J., & McGarty, C. (2002). Processes in stereotype formation: Perceivers actively make sense of groups. In C.McGarty, V.Y.Yzerbyt & R.Spears (Eds.) Stereotypes as explanations: The formation of meaningful beliefs about social groups (pp. 90–110). Cambridge: Cambridge University Press. Berscheid, E., Graziano, W., Monson, T., Dermer, M. (1976). Outcome dependency: Attention, attribution, and attraction. Journal of Personality and Social Psychology, 34, 978–989. Biernat, M., Crandall, C.S., Young, L.V., Kobrynowics, D., & Halpin, S.M. (1998). All that you can be: Stereotyping of self and others in a military context. Journal of Personality and Social Psychology, 75, 301–317. Bigler, R.S. (1995). The role of classification skill in moderating environmental effects on children’s gender stereotyping: A study of the functional use of gender in the classroom. Child Development, 66, 1072–1087. Bigler, R.S., Jones, L.C., & Lobliner, D.B. (1997). Social categorization and the formation of intergroup attitudes in children. Child Development, 68, 530–543. Billig, M. (1976). Social psychology and intergroup relations. London: Academic Press. Billig, M. (1991). Ideology and opinions. London: Sage. Billig, M. (1995). Banal nationalism. London: Sage. Blanz, M. (1999). Accessibility and fit as determinants of the salience of social categorizations. European Journal of Social Psychology, 29, 43–74. Blascovich, J., Wyer, N.A., Swart, L.A., & Kibler, J.L. (1997). Racism and racial categorization. Journal of Personality and Social Psychology, 72, 1364–1372. Blee, K.M. (2002). Inside organized racism: Women in the hate movement. Berkeley, CA: University of California Press. Bless, H., Strack, F., & Schwarz, N. (1993). The informative functions of research procedures: Bias and the logic of conversations. European Journal of Social Psychology, 23, 149–165. Bliuc, A.-M., McGarty, C., & Reynolds, K.J. (2000). The effects of intergroup conflict on social categorization and uncertainty: Distinguishing between self-categorization theory and the uncertainty reduction model. Unpublished manuscript, Australian National University. Bodenhausen, G.V., & Lichtenstein, M. (1987). Social stereotypes and information-processing strategies: The impact of task complexity. Journal of Personality and Social Psychology, 52, 871–880. Bodenhausen, G., & Macrae, C.N. (1998). Stereotype activation and inhibition. In R. S.Wyer (Ed.), Advances in social cognition (Vol. 11, pp. 1–52). Hillsdale, NJ: Erlbaum. Boen, F., & Vanbeselaere, N. (1998). Reactions upon a failed attempt to enter a high status group: An experimental test of the five-stage model. European Journal of Social Psychology, 28, 689–696. Bogart, L.M. (1998). The relationship of stereotypes about helpers to help-seeking judgments, preferences, and behaviors. Personality and Social Psychology Bulletin, 24, 1264–1275. Bogart, L., Ryan, C., & Stefanov, M. (1999). Effects of stereotypes and outcome dependency on the processing of information about group members. Group Processes and Intergroup Relations, 2, 31–50. Boldry, J.G., & Kashy, D.A. (1999). Intergroup perception in naturally occurring groups of differential status: A social relations perspective. Journal of Personality and Social Psychology, 77, 1200–1212. Borgatta, E. & Cottrell, L. (1955). On the classification of groups. Sociometry, 18, 409– 422. Bornstein, G. (1992). The free rider problem in intergroup conflict over step-level and continuous public goods. Journal of Personality and Social Psychology, 62, 597–606. Boyanowsky, E., & Allen, V. (1973). Ingroup norms and self-identity as determinants of discriminatory behavior. Journal of Personality and Social Psychology, 25, 408–418. Boyer, P. (1993). Pseudo-natural kinds. In P. Boyer (Ed.), Cognitive aspects of religious symbolism (pp. 121–141). Cambridge: Cambridge University Press. Braddock, J.H., & McPartland, J.M. (1987). Social psychology and affirmative action: The importance of process in policy analysis. Journal of Social Issues, 43, 127–132.
330
REFERENCES
Branscombe, N.R. (1998). Thinking about one’s gender group’s privileges or disadvantages: Consequences for wellbeing in women and men. British Journal of Social Psychology, 37, 167–184. Branscombe, N.R., Wann, D.L., Noel, J.G., & Coleman, J. (1993). Ingroup or outgroup extremity: Importance of the threatened social identity. Personality and Social Psychology Bulletin, 19, 381–388. Brauer, M. (2001). Intergroup perception in the social context: The effects of social status and group membership on outgroup homogeneity and ethnocentrism. Journal of Experimental Social Psychology, 37, 15–31. Brauer, M., Chamgers, P., Niedenthal, P., & Chatand-Pannetier, T. (in press). The relationship between expertise and evaluative extremity. Journal of Personality and Social Psychology. Brauer, M., & Judd, C.M. (1996). Group polarization and repeated attitude expressions. In W.Stroebe & M.Hewstone (Eds.), European Review of Social Psychology (Vol. 7, pp. 174–207). Chichester, UK: John Wiley & Sons. Brauer, M., & Judd, C.M. (2000). Defining variables in relationship to other variables: When interactions suddenly turn out to be main effects. Journal of Experimental Social Psychology, 36, 410–423. Brauer, M., Judd, C.M., & Gliner, M.D. (1995), The effects of repeated expression on attitude polarization during group discussions. Journal of Personality and Social Psychology, 68, 1014–1029. Brauer, M.. Judd, C.M., & Jacquelin, V. (2001). The communication of social stereotypes: The effects of group discussion and information distribution on stereotypic appraisals. Journal of Personality and Social Psychology, 81, 463–475. Brewer, M.B. (1979). Ingroup bias in the minimal intergroup situation: A cognitive-motivational analysis. Psychological Bulletin, 86, 307–324. Brewer, M.B. (1988). A dual process model of impression formation. In T.K.Srull & R.S. Wyer, Jr. (Eds.), Advances in social cognition (Vol. 1, pp. 1–36). Hillsdale, NJ: Erlbaum. Brewer, M.B. (1991). The social self: On being similar and different at the same time. Personality and Social Psychology Bulletin, 17, 475–482. Brewer, M.B. (1993). Social identity, distinctiveness, and in-group homogeneity. Social Cognition, 11(1), 150–164. Brewer, M.B. (2000). Reducing prejudice through cross-categorization: Effects of multiple social identities. In S.Oskamp (Ed.), Reducing prejudice and discrimination (pp. 165–183). Mahwah, NJ: Erlbaum. Brewer, M.B., & Alexander, M.G. (2002). Intergroup images and emotions. In D. Mackie & E.Smith (Eds.), From prejudice to intergroup emotions (pp. 209–225). Philadelphia: Psychology Press. Brewer, M.B., & Brown, R. (1998). Intergroup relations. In D.T.Gilbert, S.T.Fiske, & G. Lindzey (Eds.), The handbook of social psychology, (pp. 554–594). Boston: McGraw-Hill. Brewer, M.B., & Caporael, L.R. (1990). Selfish genes vs. selfish people: Sociobiology as origin myth. Motivation and Emotion, 14, 237–243. Brewer, M.B., Dull, V., & Lui, L. (1981). Perceptions of the elderly: Stereotypes and prototypes. Journal of Personality and Social Psychology, 41, 656–670. Brewer, M.B., & Harasty, A.S. (1996). Seeing groups as entities: The role of perceiver motivation. In R.Sorrentino & E.T.Higgins (Eds.), Handbook of motivation and cognition (Vol. 3, pp. 347–370). New York: Guilford Press. Brewer, M.B. & Harasty, A.S., & Feinstein, A.S. (1999). Dual processes in the cognitive representation of persons and social categories. In S.Chaiken & Y.Trope (Eds.), Dual process theories in social psychology (pp. 255–270). New York: Guilford Press. Brewer, M.B., & Lui, L. (1984). Categorization of the elderly by the elderly: Effects of perceiver’s category membership. Personality and Social Psychology Bulletin, 10, 585–595. Brewer, M.B., & Lui, L.N. (1989). The primacy of age and sex in the structure of person categories. Social Cognition, 7, 262–274. Brewer, M.B., & Weber, J.G. (1994). Self-evaluation effects of interpersonal versus intergroup social comparison. Journal of Personality and Social Psychology, 66, 268– 275. Brewer, M.B., Weber, J.G., & Carini, B. (1995). Person memory in intergroup contexts: Categorization versus individuation. Journal of Personality and Social Psychology, 69, 29–40. Brigham, J.C. (1971). Ethnic stereotypes. Psychological Bulletin, 76, 15–38.
REFERENCES
331
Brinthaupt, T., Moreland, R.L., & Levine, J. M. (1991). Sources of optimism among prospective group members. Personality and Social Psychology Bulletin, 17, 36–43. Brock, T.C., & Fromkin, H.L. (1968). Cognitive tuning set and behavioral receptivity to discrepant information. Journal of Personality, 36, 108–125. Brown, P., & Turner, J.C. (2002). The role of theories in the formation of stereotype content. In C.McGarty, V.Y.Yzerbyt, & R. Spears (Eds.), Stereotypes as explanations: The formation of meaningful beliefs about social groups (pp. 68–89). Cambridge: Cambridge University Press. Brown, R., & Smith, A. (1989). Perceptions of and by minority groups: The case of women in academia. European Journal of Social Psychology, 19, 61–75. Brown, R., & Turner, J.C. (1981). Interpersonal and intergroup behaviour. In J.C. Turner & H.Giles (Eds.), Intergroup behaviour (pp. 33–65). Oxford: Blackwell. Brown, R., & Wootton-Millward, L. (1993). Perceptions of group homogeneity during group formation and change. Social Cognition, 11, 126–149. Bruner, J.S. (1957). On perceptual readiness. Psychological Review, 64, 123–152. Burnstein, E., & Vinokur, A. (1975). What a person thinks on learning he has chosen differently from others: Nice evidence for the persuasive arguments explanation of choice shifts. Journal of Experimental Social Psychology, 11, 412–426 Burnstein, E., & Vinokur, A. (1977). Persuasive argumentation and social comparison as determinants of attitude polarization. Journal of Experimental Social Psychology, 13, 315–332. Cabecinhas, R., & Amâncio, L. (1999). Asymmetries in the perception of other as a function of social position and context. Swiss Journal of Psychology, 58, 40–50. Cacioppo, J.T., & Petty, R.E. (1982). The need for cognition. Journal of Personality and Social Psychology, 42, 116–131. Cadinu, M.R., & Rothbart, M. (1996). Self-anchoring and differentiation processes in the minimal group setting. Journal of Personality and Social Psychology, 70, 661–677. Campbell, D.T. (1958) Common fate, similarity, and other indices of the status of aggregates of persons as social entities. Behavioural Science, 3, 14–25. Campbell, D.T. (1965). Ethnocentrism and other altruistic motives. In D.Levine (Ed.), Nebraska symposium on motivation (Vol. 13, pp. 283–311). Lincoln: University of Nebraska Press. Campbell, D.T. (1967). Stereotypes and perception of group differences. American Psychologist, 22, 817–829. Campbell, D.T. (1987). Evolutionary epistemology. In G.Radnitzky & W.W.Hartley, III (Eds.), Evolutionary epistemology, theory of rationality, and the sociology of knowledge (pp. 47–89). LaSalle, IL: Open Court Press. Campbell, D.T. & Fiske, D.W. (1959). Convergent and discriminant validation by the multitrait-multimethod matrix. Psychological Bulletin, 56, 81–105. Campbell, J.D. (1990). Self-esteem and the clarity of the self-concept. Journal of Personality and Social Psychology, 59, 538–549. Campbell, J.D., Trapnell, P.D., Heine, S.J., Katz, I.M., Lavalle, L.F., & Lehman, D.R. (1996). Self-concept clarity: Measurement, personality correlates, and cultural boundaries. Journal of Personality and Social Psychology, 70, 141–156. Caporael, L.R. (1997). The evolution of truly social cognition: The core configurations model. Personality and Social Psychology Review, 1, 276–298. Caporael, L.R., & Baron, R.M. (1997). Groups as the minds natural environment. In J.A. Simpson & D.T.Kenrick (Eds.), Evolutionary social psychology (pp. 317–344). Hillsdale, NJ: Erlbaum. Carlston, D.E., & Skowronski, J.J. (1994). Savings in the relearning of trait information as evidence for spontaneous inference generation. Journal of Personality and Social Psychology, 66, 840–856. Carver, C.S., & Scheier, M.F. (1981). Attention and self-regulation: A control theory approach to human behavior. New York: Springer-Verlag. Carver, C.S., & Scheier, M.F. (1998). On the self-regulation of behavior. Cambridge: Cambridge University Press.
332
REFERENCES
Castano, E. (1999). The phenomenology of the ingroup: Entitativity and identification. Unpublished doctoral dissertation, Catholic University of Louvain at Louvain-la-Neuve. Castano, E. (2002). The European Union and Canada: Laboratoires Identitaires a hauts risques. Paper presented at the Thinkers’ Retreat. Canada-Europe: Futures for Trans-Atlantic Relations (consultant). Ministry of Foreign Affairs and International Trade, Ottawa, Canada. Castano, E. (2003, January). Ethnos, demos and project. Is it possible to construct a common identity even in the absence of a European “people”? European Synthesis. Castano, E. (in press). European identity: A social psychological perspective. In R.H. Herrmann, M.B., Brewer, & T.Risse (Eds.), Identities in Europe and the institutions of the European Union. Lanham, MD: Rowman and Littlefield. Castano, E., & Brewer, M.B. (2000). Entitativity in artificial groups. Unpublished manuscript, Ohio State University. Castano, E., Brewer, M.B., & MacDonald, K. (2000). How badly off we are! Negative common fate, entitativity and identification with the ingroup. Unpublished manuscript, Ohio State University. Castano, E., Paladino, M.P., Coull, A., & Yzerbyt, V.Y. (1999). Preserving the ingroup entitativity: Ingroup identification and the management of deviant ingroup members. Manuscript submitted for publication. Catholic University of Louvain at Louvainla-Neuve, Belgium. Castano, E., Paladino, P., Coull, A., & Yzerbyt, V.Y. (in press). Dealing with a deviant ingroup member. British Journal of Social Psychology. Castano, E., & Sacchi, S. (2000). From bad to worse, from good to better: The polarizing effect of entitativity. Unpublished manuscript, Ohio State University. Castano, E., Sacchi, S., Gries, P.H. (2003). The perception of “The Other” in international relations: Evidence for the polarizing effect of entitativity. Political Psychology, 24, 449– 468. Castano E., & Tousignant, N., (1999). La Belgique et l’Europe. Un demos sans ethnos? In G.-F.Dumont (Ed), Les racines de l’identité europeenne (pp. 99–123), Paris: Economica Paris/Thesis Zuri. Castano, E., & Yzerbyt, V.Y. (1998). The highs and lows of group homogeneity. Behavioral processes, 42, 219–238. Castano, E., Yzerbyt, V.Y., & Bourguignon, D. (1999). Measuring entitativity. Unpublished Manuscript. Catholic University of Louvain at Louvain-la-Neuve. Castano, E., Yzerbyt, V.Y., Bourguignon, D. (in press). We are one and I like it: The impact of ingroup entitativity on social identification. European Journal of Social Psychology. Castano, E., Yzerbyt, V.Y., Bourguignon, D., & Seron, E. (2002). Who may come in? The impact of ingroup identification on ingroup-outgroup categorization. Journal of Experimental Social Psychology, 38, 315–322. Castano, E., Yzerbyt, V.Y., Paladino, M.P. (in press). Attaining immortality through social identification. In J.Greenberg, S.L.Koole, & T.Pyszczynski (Eds.), Handbook of experimental existential psychology. New York: Guilford Press. Castano, E., Yzerbyt, V.Y., Paladino, M.P., & Sacchi, S. (2002). I belong therefore I exist: Intragroup dynamics and the management of the anxiety of death. Personality and Social Psychology Bulletin, 28, 135–143. Castano, E., Yzerbyt, V.Y., & Tousignant, N. (2000). Europeans become European: Identity management in today’s Europe. Manuscript in preparation. Cattell, R.B. (1948). Concepts and methods in the measurement of group syntality. Psychological Review, 55, 48–63. Chao, G.T., O’Leary-Kelly, A.M., Wolf, S., Klein, H.J., & Gardner, P.D. (1994). Organizational socialization: Its content and consequences. Journal of Applied Psychology, 79, 730–743. Chapman, L.J., & Chapman, J. (1967). Genesis of popular but erroneous psychodiagnostic observations. Journal of Abnormal Psychology, 72, 193–204. Chartrand, T.L., & Bargh, J.A. (1999). The chameleon effect: The perception-behavior link and social interaction. Journal of Personality and Social Psychology, 76, 893–910. Chen, S., Lee-Chai, A.Y, & Bargh, J. (2001). Relationship orientation as a moderator of the effects of social power. Journal of Personality and Social Psychology, 80, 173–187. Chiu, C. (1994). Bases of categorization and person cognition. Unpublished doctoral dissertation, Columbia University.
REFERENCES
333
Chiu, C., Dweck, C.S., Tong, Y.Y., & Fu, H. Y. (1997). Implicit theories and conceptions of morality. Journal of Personality and Social Psychology, 73, 923–940. Chiu, C., & Hong, Y. (1999). Social identification in a political transition: The role of implicit beliefs. International Journal of Inter-cultural Relations, 23, 297–318. Chiu, C., Hong, Y., & Dweck, C.S. (1997). Lay dispositionism and implicit theories of personality. Journal of Personality and Social Psychology, 73, 19–30. Chiu, C., Morris, M.W., Hong, Y., & Menon, T. (2000). Motivated cultural cognition: The impact of lay cultural theories on dispositional attribution varies as a function of need for closure. Journal of Personality and Social Psychology, 78, 247–259. Chow, A.P.Y. (1996). Lay theories and inter-group perception. Unpublished undergraduate thesis, University of Hong Kong. Cialdini, R.B. (1989). Indirect tactics of impression management: Beyond basking. In R.A.Giacalone & P.Rosenfeld (Eds.), Impression management in the organization (pp. 45–56). Hillsdale, NJ: Erlbaum. Cialdini, R.B., Borden, R.J., Thorne, A., Walker, M.R., Freeman, S., & Sloan, L.R. (1976). Basking in reflected glory: Three (football) field studies. Journal of Personality and Social Psychology, 34, 366–375. Clark, H.H. (1996). Using language. Cambridge: Cambridge University Press. Clark, H.H., & Brennan, S.E. (1991). Grounding in communication. In L.B.Resnick, J. M.Levine, & S.D.Teasly (Eds.), Perspectives on socially shared cognition (pp. 127– 149). Washington, DC: American Psychological Association. Clark, H.H., & Wilks-Gibbs, D. (1986). Referring as a collaborative process. Cognition, 22, 1–39. Clemen, R.T., Fischer, G.W., and Winkler, R. L. (2000). Assessing dependence: Some experimental results. Management Science, 46, 1100–1115, Codol, J.-P. (1984). Social differentiation and non-differentiation. In H.Tajfel (Ed.), The social dimension: European developments in social psychology (Vol. 1, pp. 314–337). Cambridge: Cambridge University Press. Cohen, A.R. (1961). Cognitive tuning as a factor affecting impression formation. Journal of Personality, 29, 235–245. Coovert, M.D., & Reeder, G.D. (1990). Negativity effects in impression formation: The role of unit formation and schematic expectations. Journal of Experimental Social Psychology, 26, 49–62. Corneille, O., & Judd, C.M. (1999). Accentuation and sensitization effects in the categorization of multifaceted stimuli. Journal of Personality and Social Psychology, 77, 927–941. Corneille, O., Klein, O., Lambert, S., & Judd, C.M. (2002). On the role of familiarity with units of measurement in categorical accentuation: Tajfel and Wilkes (1963) revisited and replicated. Psychological Science, 4, 380–383. Corneille, O., Vescio, T.K., & Judd, C.M. (2000). Incidentally activated knowledge and stereotype based judgments: A consideration of primed construct—target attribute match. Social Cognition, 18, 377– 399. Corneille, O., & Yzerbyt, V. (2002): Dependence and the formation of stereotyped beliefs about groups: From interpersonal to intergroup perception. In C.McGarty, V. Yzerbyt & R.Spears (Eds.), Stereotypes as explanations: The formation of meaningful beliefs about social groups (pp. 111–126). Cambridge, UK: Cambridge University Press. Corneille, O., Yzerbyt, V.Y., Rogier, A., & Buidin, G. (2001). Threat and the group attribution error: When threat elicits judgments of extremity and homogeneity. Personality and Social Psychology Bulletin, 27, 437–446. Coull, A., Yzerbyt, V.Y., Castano, E., Paladino, M., & Leemans, V. (2001). Protecting the ingroup: Motivated allocation of cognitive resources in the presence of threatening ingroup members. Group Processes and Intergroup Relations, 4, 327–339, Crawford, M.T., Sherman, S.J., & Hamilton, D.L. (2002). Perceived entitativity, stereotype formation, and the interchangeability of group members. Journal of Personality and Social Psychology, 83, 1076–1094. Crenshaw, M. (1985). An organizational approach to the analysis of political terrorism. Orbis, 29, 465–489. Crisp, R.J., & Hewstone, M. (1999). Differential evaluation of crossed category Patterns, processes, and reducing intergroup bias, Group Processes and Intergroup Relations, 2, 307–333. Crocker, J. (1981). Judgment of covariation by social perceivers. Psychological Bulletin, 90, 272–292. Crocker, J., Fiske, S.T., & Taylor, S.E. (1984). Schematic basis of belief change. In J.R. Eiser (Ed.), Attitudinal judgment (pp. 197– 226). New York: Springer-Verlag.
334
REFERENCES
Crocker, J., & Major, B. (1989). Social stigma and self-esteem: The self-protective properties of stigma. Psychological Review, 96, 608–630. Crocker, J., Major, B., & Steele, C. (1998). Social stigma. In D.T.Gilbert, S.T.Fiske, & G.Lindzey (Eds.), The handbook of social psychology (4th ed., Vol. 2, pp. 504– 553). Boston: McGraw-Hill. Crockett, W.H. (1965). Cognitive complexity and impression formation. In B.A.Maher (ed.), Progress in experimental personality research (Vol. 2, pp. 47–90). New York: Academic Press. Curtis, J.M., & Curtis, M.J. (1993). Factors related to susceptibility and recruitment by cults. Psychological Reports, 73, 451–460. D’Andrade, R.G. (1974). Memory and the assessment of behavior. In H.Blalock (Ed.), Measurement in the social sciences. Chicago: Aldine, 159–186. D’Andrade, R.G. (1984). Cultural meaning systems. In R.A.Shweder & R.A.LeVine (Eds.). Culture theory (pp. 88–119). Cambridge: Cambridge University Press. Dasgupta, N., Banaji, M.R., & Abelson, R.P. (1999). Group entitativity and group perception: Associations between physical features and psychological judgment. Journal of Per sonality and Social Psychology, 77, 991– 1003. Davies, P. (1992). The role of disclosure in coming out among gay men. In K.Plummer (Ed.), Modern homosexualities: Fragments of lesbian and gay experience (pp. 75–83). London: Routledge. Davis, J., & Witte, E. (1996). Understanding group behavior: Consensual action by small groups (Vol. 2). Mahwah, NJ: Erlbaum. Deaux, K., & Lewis, L.L. (1984). Structure of gender stereotypes: Interrelationships among components and gender labels. Journal of Personality and Social Psychology, 5, 991–1004. Dechesne, M., Greenberg, J., Arndt, J., & Schimel, J. (2000). Terror management and the vicissitudes of sports fan affiliation: The effects of mortality salience on optimism and fan identification. European Journal of Social Psychology, 30, 813–835. Dechesne, M., Janssen, J., van Knippenberg, A. (2000). Derogation and distancing as terror management strategies: The moderating role of need for closure and permeability of group boundaries. Journal of Personality and Social Psychology, 79, 923–932. Deci, E.L., & Ryan, R.M. (1991). A motivational approach to self: Integration in personality. In R.A.Dienstbier (Ed.), Nebraska Symposium on Motivation: Perspectives on motivation (pp. 237–288). Lincoln: University of Nebraska Press. Della Porta, D. (1995). The logic of underground organizations: Social movements, political violence, and the state. Cambridge: Cambridge University Press. Della Porta, D., & Diani, M. (1999). Social movements: An introduction. Maiden, Mass: Blackwell. DeNardo, J. (1985). Power in numbers. The political strategy of protest and rebellion. Princeton, NJ: Princeton University Press. Dépret, E., & Fiske, S.T. (1999). Perceiving the powerful: Intriguing individuals versus threatening groups. Journal of Experimental Social Psychology, 35, 461–480. Deschamps, J.C., & Doise, W. (1978). Crossed category memberships in intergroup relations. In H.Tajfel (Ed.), Differentiation between social groups (pp. 142–158). London: Academic Press. Desforges, D.M., Lord, C.G., & Pugh, M.A. (1997). Role of group representativeness in the generalization part of the contact hypothesis. Basic and Applied Social Psychology, 19, 183–204. Deutsch, M., & Gerard, H.B. (1955). A study of normative and informational social influences upon individual judgment. Journal of Abnormal and Social Psychology, 51, 629– 636. Devine, P.G. (1989). Stereotypes and prejudice: Their automatic and controlled components. Journal of Personality and Social Psychology, 56, 5–18. Devine, P.G., & Baker, S.M. (1991). Measurement of racial stereotype subtyping. Personality and Social Psychology Bulletin, 17, 44– 50. Devos, T., Comby, L., & Deschamps, J. (1996). Asymmetries in judgements of ingroup and outgroup variability. In W.Stroebe & M. Hewstone (Eds.), European review of social psychology (Vol. 7, pp. 96–144). Chichester, UK: Wiley.
REFERENCES
335
Dhawan, N., Roseman, I.J., Naidu, R.K., Thapa, K., & Rettek, S.I. (1995). Self-concepts across two cultures: India and the United States. Journal of Cross-Cultural Psychology, 26, 606–621. Diehl, M. (1990). The minimal group paradigm: Theoretical explanations and empirical findings. European Review of Social Psychology, 1, 263–292. Dijksterhuis, A., Bargh, J., & Miedema, A. (2000). Of men and mackerels: Attention, subjective experience, and automatic social behavior. In H.Bless & J.Forgas (Eds.) The message within: The role of subjective experience in social cognition and behavior (pp. 37–51). Philadelphia: Psychology Press. Dijksterhuis, A., & van Knippenberg, A. (2000). Behavioral indecision: effects of self-focus on automatic behavior. Social Cognition, 18, 55–74. Dillard, J.P., & Miller, K.I. (1988). Intimate relationships in task environments. In S.W. Duck (Ed.), Handbook of personal relationships (pp. 449–465). New York: Wiley. Ditto, P.H., & Jemmott, J.B.I. (1989). From rarity to evaluative extremity: Effects of prevalence information on evaluations of positive and negative characteristics. Jour nal of Personality and Social Psychology, 57, 16–26. Doise, W. (1969). Intergroup relations and polarization of individual and collective judgments. Journal of Personality and Social Psychology, 12, 136–143. Doosje, B., Ellemers, N., & Spears, R. (1995). Perceived intragroup variability as a function of group status and identification. Journal of Experimental Social Psychology, 31, 410–436. Doosje, B., Spears, R., & Ellemers, N. (2002). Social identity as both cause and effect: The development of group identification in response to anticipated and actual changes in the intergroup status hierarchy. British Journal of Social Psychology, 41, 57–76. Doosje, B., Spears, R., Ellemers, N., & Koomen, W. (1999). Perceived group variability in intergroup relations: The distinctive role of social identity. In W.Stroebe & M.Hewstone (Eds.), European review of social psychology (Vol. 10, pp. 41–74). New York: Wiley. Doosje, B., Spears, R., & Koomen, W. (1995). When bad isn’t all bad: Strategic use of sample information in generalization and stereotyping. Journal of Personality and Social Psychology, 69, 642–655. Dovidio, J., Ellyson, S., Keating, C., & Heltman, K. (1988). The relationship of social power to visual displays of dominance between men and women. Journal of Personality and Social Psychology, 54, 233–242. Dovidio, J.F., & Gaertner, S.L. (2000). Aversive racism and selection decisions: 1989 and 1999. Psychological Science, 11, 315–319. Dovidio, J.F., Isen, A.M., Guerra, P., Gaertner, S.L., & Rust, M. (1998). Positive affect, cognition, and the reduction of intergroup bias. In C.Sedikides, J.Schopler, & C.A. Insko (Eds.), Intergroup cognition and intergroup behavior (pp. 337–366). Mahwah, NJ: Erlbaum. Downing, J.M., Judd, C.M., & Brauer, M. (1992). Effects of repeated expressions on attitude extremity. Journal of Personality and Social Psychology, 63, 17–29. Dunbar, R.I.M. (1996). Grooming, gossip, and the evolution of language. London: Faber & Faber. Dunn, R.G. (1998). Identity crises: A social critique of postmodernity. Minneapolis: University of Minnesota Press. Dupré, J. (1981). Natural kinds and biological taxa. Philosophical Review, 90, 66–90. Durkheim, E. (1893/1984: reprinted 1993). The division of labor in society. London: Macmillan. Durkheim, E. (1901/1982). The rules of sociological method. [Ed. by S.Lukes & Tr. by W. D.Halls]. London: Macmillan. Dutton, J.E., Dukerich, J.M., & Harquail, C. V. (1994). Organizational images and member identification. Administrative Science Quarterly, 39, 239–263. Dweck, C.S. (1999). Self-theories: Their role in motivation, personality, and development. Philadelphia: Psychology Press. Dweck, C.S., Chiu, C., & Hong, Y. (1995). Implicit theories and their role in judgments and reactions: A world from two perspectives. Psychological Inquiry, 6, 267–285. Dweck, C.S., & Leggett, E.L. (1988). A social-cognitive approach to motivation and personality. Psychological Review, 25, 109– 116. Eagly, A.H. (1987). Sex differences in social behavior: A social-role interpretation. Hillsdale, NJ: Erlbaum.
336
REFERENCES
Eagly, A.H., & Chaiken, S. (1993). The psychology of attitudes. San Diego, CA: Harcourt Brace Jovanovich. Eagly, A.H., Mladinic, A., & Otto, S. (1994). Cognitive and affective bases of attitudes toward social groups and social policies. Journal of Experimental Social Psychology, 30, 113–137. Eagly, A.H., & Wood, W. (1982). Inferred sex differences in status as a determinant of gender stereotypes about social influence. Journal of Personality and Social Psychology, 43, 915–928. Earle, W.B. (1986). The social context of social comparison: Reality versus reassurance. Personality and Social Psychology Bulletin, 12, 159–168. Ebenbach, D.H., & Keltner, D. (1998). Power, emotion, and judgmental accuracy in social conflict: Motivating the cognitive miser. Basic and Applied Social Psychology, 20, 7– 21. Einhorn, H.J. (1982). Learning from experience and suboptimal rules in decision mak ing. In D.Kahneman, P.Slovic, and A. Tversky (Eds.), Judgment under uncertainty: Heuristics and biases (pp. 268–283), Cambridge: Cambridge University Press. Ellemers, N. (1993). The influence of socio-structural variables on identity enhancement strategies. In W.Stroebe and M.Hewstone (Eds.), European review of social psychology (Vol. 4, pp. 27–57). Chichester, UK: Wiley. Ellemers, N., Doosje, B., van Knippenberg, A., & Wilke, H. (1992) Status protection in high status minority groups. European Journal of Social Psychology, 22, 123–40. Ellemers, N., Spears, R., & Doosje, B. (1997). Sticking together or falling apart: Ingroup identification as a psychological determinant of group commitment versus individual mobility. Journal of Personality and Social Psychology, 72, 617–626. Ellemers, N., van Knippenberg, A., de Vries, N.K., & Wilke, H., (1988). Social identification and permeability of group boundaries. European Journal of Social Psychology, 18, 497–513. Ellemers, N., & Van Rijswijk, W. (1997). Identity needs versus social opportunities: The use of group-level and individual-level identity management strategies. Social Psychology Quarterly, 60, 52–65. Epstein, S. (1989). Values from the perspective of cognitive-experiential self-theory. In N.Eisenberg, J.Reykowski, & E.Staub (Eds.), Social and moral values: Individual and social perspectives, (pp. 3–61). Hillsdale, NJ: Erlbaum. Erber, R., & Fiske, S.T. (1984). Outcome dependency and attention to inconsistent information. Journal of Personality and Social Psychology, 47, 709–726. Erdley, C.A., & Dweck, C.S. (1993). Children’s lay personality theories as predictors of their social judgments. Child Development, 64, 863–878. Esses, V.M., Haddock, G., & Zanna, M.P. (1993). Values, stereotypes, and emotions as determinants of intergroup attitudes. In D. M.Mackie & D.L.Hamilton (Eds.), Affect, cognition, and stereotyping: Interactive processes in group perception (pp. 137–166). San Diego, CA: Academic Press. Estrada, C., & Yzerbyt, V.Y. (2002). Unpublished data. Catholic University of Louvain at Louvain-la-Neuve. Fábrega, H. (1997). The evolution of sickness and healing. Berkeley, CA: University of California Press. Falcione, R.L., & Wilson, C.E. (1988). Socialization processes in organizations. In G. M.Goldhaber & G.A.Barnett (Eds.), Handbook of organizational communication (pp. 151–169). Norwood, NJ: Ablex. Farley, J. (1982). Majority-minority relations. Englewood Cliffs, NJ: Prentice Hall. Festinger, L. (1950). Informal social communication. Psychological Review, 57, 271–282. Festinger, L. (1954a). A theory of social comparison processes. Human Relations, 7, 117– 140. Festinger, L. (1954b). Motivation leading to social behavior. In M.R.Jones (Ed.), Nebraska symposium on motivation (vol. 2, pp. 121–218). Lincoln, NE: University of Nebraska Press. Festinger, L. (1957/1962). A theory of cognitive dissonance. Stanford, CA: Stanford University Press. Fichman, M., & Levinthal, D.A. (1991). Honeymoons and the liability of adolescence: A new perspective on duration dependence in social and organizational relationships. Academy of Management Review, 16, 442–468. Fiedler, K. (1991). The tricky nature of skewed frequency tables: An information loss account of distinctiveness-based illusory correlations. Journal of Personality and Social Psychology, 60, 24–36. Fiedler, K. (1996). Explaining and simulating judgment biases as an aggregation phenomenon in probabilistic, multiplecue environments. Psychological Review, 103, 193–214.
REFERENCES
337
Fiedler, K., Semin, G., & Finkenauer, C. (1993). The battle of words between gender groups: A language-based approach to intergroup processes. Human Communication Research, 19, 409–441. Fiske, S.T. (1993). Controlling other people: The impact of power on stereotyping. American Psychologist, 48, 621–628. Fiske, S.T. (1998). Stereotyping, prejudice, and discrimination. In D.Gilbert, S.T. Fiske, & G.Lindzey (Eds.), Handbook of social psychology (4th ed., Vol. 2, pp. 357– 411). New York: McGraw-Hill. Fiske, S.T. (2000). Interdependence reduces prejudice and stereotyping. In S.Oskamp (Ed.), Reducing prejudice and discrimination (pp. 115–135). Mahwah, NJ: Erlbaum. Fiske, S.T., & Dépret, E. (1996). Control interdependence and power: Understanding social cognition in its social context. In W. Stroebe & M.Hewstone (Eds.), European review of social psychology (vol 7, pp. 31– 61). Chichester, UK: Wiley. Fiske, S.T., & Goodwin, S.A. (1996). Social cognition research and small group research: A West Side Story or…? In J.L.Nye & A. M.Brower (Eds.), What’s social about social cognition? Research on socially shared cognition in small groups. Thousand Oaks, CA: Sage. Fiske, S.T., Morling, B.A., & Stevens, L.E. (1996). Controlling self and others: A theory of anxiety, mental control, and social control. Personality and Social Psychology Bulletin, 22, 115–123. Fiske, S.T., & Neuberg, S.L. (1990) A continuum of impression formation, from category-based to individuating processes: Influences of information and motivation on attention and interpretation. In M.P.Zanna (Ed.), Advances in experimental social psychology (Vol. 23, pp. 1–73) New York: Random House. Fiske, S.T., & Taylor, S.E. (1991). Social cognition (2nd ed.). New York: McGraw-Hill. Fletcher, G.J.O., Danilovics, P., Fernandez, G., Peterson, D., & Reeder, G.D. (1986). Attributional complexity: An individual differences measure. Journal of Personality and Social Psychology, 51, 875–84. Ford, T.E., & Stangor, C. (1992). The role of diagnosticity in stereotype formation: Perceiving group means and variances. Journal of Personality and Social Psychology, 63, 356–67. Förster, J., Higgins, E.T., & Strack, F. (2000). When stereotype disconfirmation is a personal threat: How prejudice and prevention focus moderate incongruency effects. Social Cognition, 18, 178–197. Forsyth, D.R. (1990). Group dynamics. Pacific Grove, CA: Brooks/Cole Publishing Company. Frable, D.E.S., & Bem, S.L. (1985). If you are gender schematic, all members of the opposite sex look alike. Journal of Personality and Social Psychology, 49, 459–468. Freeman, L.C. (1992). Filling in the blanks: A theory of cognitive categories and the structure of social affiliation. Social Psychology Quarterly, 55, 118–127. French, J.R.P., Jr., & Raven, B.H. (1959). The bases of social power. In D.Cartwright (Ed.), Studies in social power (pp. 150–167). Ann Arbor, MI: Institute for Social Research. Fuss, D. (1989). Essentially speaking: Feminism, nature, and difference. New York: Routledge. Gaertner, L., & Insko, C.A. (2000). Intergroup discrimination in the minimal group paradigm: Categorization, reciprocation, or fear? Journal of Personality and Social Psychology, 79, 77–94. Gaertner, L., & Schopler, J. (1998). Perceived ingroup entitativity and intergroup bias: An interconnection of self and others. European Journal of Social Psychology, 28, 963–980. Gaertner, L., Sedikides, C., & Graetz, K. (1999). In search of self-definition: Motivational primacy of the individual self, motivational primacy of the collective self, or contextual primacy? Journal of Personality and Social Psychology, 76, 5–18. Gaertner, S.L., Dovidio, J. F., Anastasio, P.A., Bachman, B.A., & Rust, M.C. (1993). The Common Ingroup Identity Model: Recategorization and the reduction of intergroup bias. In W.Stroebe & M.Hewstone (Eds.), European review of social psychology (Vol. 4, pp. 1–26). Chichester, UK: Wiley. Gaertner, S.L., Dovidio, J. F., Nier, J.A., Ward, C.M., & Banker, B.S. (1999). Across cultural divides: The value of a superordinate identity. In D.Prentice & D.Miller (Eds.), Cultural Divides: Understanding and overcoming group conflict (pp. 173–212). New York: Russell Sage Foundation. Gaertner, S.L., Dovidio, J. F., Rust, M.C., Nier, J.A., Banker, B.S., Ward, C.M., et al.. (1999). Reducing intergroup bias: Elements of intergroup cooperation. Journal of Personality and Social Psychology, 76, 388–402.
338
REFERENCES
Gaertner, S.L., Mann, J.A., Murrell, A., & Dovidio, J.F. (1989). Reducing intergroup bias: The benefits of recategorization. Journal of Personality and Social Psychology, 57, 239–249. Galanter, M. (Ed.) (1989). Cults and new religious movements. Washington, DC: American Psychiatric Association. Gamson, W.A. (1995). Hiroshima, the Holocaust, and the politics of exclusion. American Sociological Review, 60, 1–20. Garcia-Marques, L., & Mackie, D.M. (1999). The impact of stereotype-incongruent information on perceived group variability and stereotype change. Journal of Personality and Social Psychology, 77, 979–990. Geis, F.L. (1993). Self-fulfilling prophecies: A social psychological view of gender. In A. E.Beall & R.J.Sternberg (Eds.), The psychology of gender (pp. 9–54). New York: Guilford Press. Gelman, S.A. (1988). The development of induction within natural kind and artifact categories. Cognitive Psychology, 20, 65–95. Gelman, S.A., & Coley, J.D. (1991). Language and categorization: The acquisition of natural kind terms. In Gelman, S.A., & Byrnes, J.P. (Eds.), Perspectives on language and thought: Interrelations in development (pp. 146–196). Cambridge: Cambridge University Press. Gelman, S.A., & Heyman, G.D. (1999). Carrot-eaters and creature believers: The effects of lexicalization on children’s inferences about social categories. Psychological Science, 10, 489–493. Gelman, S.A., & Hirschfeld, L.A. (1999). How biological is essentialism? In D.Medin & S. Atran (Eds.), Folkbiology. Cambridge, MA: MIT Press. Georgesen, J.C., & Harris, M. (2000). The balance of power: Interpersonal consequences of differential power expectancies. Personality and Social Psychology Bulletin, 26, 1239–1257. Gergen, K.J. (1973). Social psychology as history. Journal of Personality and Social Psychology, 26, 309–320. Gergen, K.J. (1985). The social construction-ist movement in modern psychology. American Psychologist, 40, 266–275. Gergen, K.J. (1991). The saturated self: Dilemmas of identity in contemporary life. New York: Basic Books. Gergen, K.J. (1994). Toward transformation in social knowledge (2nd ed.). London: Sage. Gervey, B.M., Chiu, C.Y., Hong, H., & Dweck, C.S. (1999). Differential use of person information in decisions about guilt versus innocence: The role of lay theories. Personality and Social Psychology Bulletin, 25, 17– 27. Gibson, J.J. (1977) The theory of affordances. In R.Shaw & J.Bransford (Eds.), Perceiving, acting and knowing. Hillsdale NJ: Erlbaum. Giddens, A. (1976/1993). New rules of sociological method (2nd ed.). Stanford, CA: Stanford University Press. Gigone, D., & Hastie, R. (1993). The common knowledge effect: Information sharing and group judgment. Journal of Personality and Social Psychology, 65, 959–974. Gigone, D., & Hastie, R. (1997). The impact of information on small group choice. Journal of Personality and Social Psychology, 72, 132–140. Gilbert, D.T. & Hixon, J.G. (1991). The trouble of thinking: Activation and application of stereotypic beliefs. Journal of Personality and Social Psychology, 60, 509–517. Giles, H., Coupland, N., & Coupland, J. (1991). Accommodation theory: Communication, context, and consequence. In H.Giles, N. Coupland, & J.Coupland (Eds), Contexts of accommodation: Developments in applied sociolinguistics. Studies in emotion and social interaction (pp. 1–68). New York: Cambridge University Press Gillespie, J.B. (1990). The phenomenon of the public wife: An exercise in Goffman’s impression management. In D.Brissett & C. Edgley (Eds.), Life as theater: A dramaturgical sourcebook (pp. 379–397). Hawthorne, NY: Aldine de Gruyter. Gil-White, F.J. (2001). Are ethnic groups biological “species” to the human brain? Essentialism in our cognition of some social categories. Current Anthropology, 42, 1–56. Glick, P., & Fiske, S.T. (1997). Hostile and benevolent sexism: Measuring ambivalent sexist attitudes toward women. Psychology of Women Quarterly, 21, 119–135. Goldsmith, D., & Parks, M.R. (1990). Communicative strategies for managing the risks of seeking social support. In S.Duck & R. C.Silver (Eds.), Personal relationships and social support (pp. 104–121). Thousand Oaks, CA: Sage.
REFERENCES
339
Goodwin, S.A., Gubin, A., Fiske, S.T., & Yzerbyt, V. (1999). Power can bias impression formation: Stereotyping subordinates by default and by design. Manuscript submitted for publication. Gramzow, R.H., Gaertner, L., & Sedikides, C. (2001). Memory for in-group and out-group information in a minimal group context: The self as an informational base. Journal of Personality and Social Psychology, 80, 188–205. Green, G.M. (1996). Pragmatics and natural language understanding (2nd ed.). Mahwah, NJ: Erlbaum. Greenberg, J., Pyszczynski, T., & Solomon, S. (1986). The causes and consequences of the need for self-esteem: A terror management theory. In R.F.Baumeister (Ed.), Public self and private self (pp. 198–212). New York: Springer-Verlag. Greenberg, J., Solomon, S., & Pyszczynski, T., (1997). Terror management theory of self-esteem and cultural worldviews: Empirical assessments and conceptual refinements. In P.M.Zanna (Ed.), Advances in experimental social psychology (Vol. 29, pp. 61–139). San Diego, Academic Press. Greenhalgh, L. (1983). Organizational decline. In S.B.Bacharach (Ed.), Research in the sociology of organizations (Vol. 2, pp. 231– 276). Greenwich, CT: JAI Press. Greenwald, A.G., & Banaji, M.R. (1995). Implicit social cognition: Attitudes, self-esteem, and stereotypes. Psychological Review, 102, 4–27. Greenwald, A.G., McGhee, D.E., & Schwartz, J.L.K. (1998). Measuring individual differences in implicit cognition: The implicit association test. Journal of Personality and Social Psychology, 74, 1464–1480. Grieve, P., & Hogg, M.A. (1999). Subjective uncertainty and intergroup discrimination in the minimal group situation. Personality and Social Psychology Bulletin, 25, 926–940. Grosz, E. (1990). A note on essentialism and difference. In S.Gunew (ed.), Feminist knowledge as critique and construct. London: Routledge. Gruenfield, D.H., & Fan, E.T. (1999). What newcomers see and what oldtimers say: Discontinuities in knowledge exchange. In L.L.Thompson, J.M.Levine, & D.M. Messick (Eds.), Shared cognition in organi zations: The management of knowledge (pp. 245–265). Mahwah, NJ: Erlbaum. Gudykunst, W.B., & Shapiro, R.B. (1996). Communication in everyday interpersonal and intergroup encounters. International Journal of Intercultural Relations, 20, 19– 46. Gudykunst, W.B., Yang, S.M., & Nishida, T. (1985). A cross-cultural test of uncertainty reduction theory: Comparisons of acquaintance, friend, and dating relationships in Japan, Korea, and the United States. Human Communication Research, 11, 407–455. Guerin, B., & Innes, J.M. (1989). Cognitive tuning sets: Anticipating the consequence of communication. Current Psychology: Research & Reviews, 8, 234–249. Guimond, S. (2000). Group socialization and prejudice: The social transmission of intergroup attitudes and beliefs. European Journal of Social Psychology, 30, 335–354. Guinote, A. (2001). The perception of group variability in a non-minority and a minority context: When adaptation leads to out-group differentiation. British Journal of Social Psychology, 40, 117–132. Guinote, A., Brown, M., & Fiske, S.T. (2000). Perceived control, processing effort and perceived group variability as a function of group size. Unpublished manuscript, University of Essex, UK. Guinote, A., & Fiske, S.T. (1999). Outcome dependency affects perceived group variability: Evidence for outgroup differentiation. Unpublished Manuscript, University of Essex, UK. Guinote, A., Judd, C.M., & Brauer, M. (2002). Effects of power on perceived and objective group variability: Evidence that more powerful groups are more variable. Journal of Personality and Social Psychology, 82, 708–721. Guinote, A., Mouro, C., Pereira, H., & Monteiro, M. (2002). Children’s perceptions of group variability as a function of status and power. Manuscript submitted for publication. Guinote, A., Trope, Y., & Murinho, C. (2003). Power, group size, and objective group variability: A cross-situational perspective. Manuscript in preparation. Guze, S.B. (1992). Why psychiatry is a branch of medicine. New York: Oxford University Press. Haas, E. (1958). The uniting of Europe: Political, social, and economic forces 1950–1957. London: Stevens. Haddad, Y.S. (1982). The effect of information versus controlling verbal feedback on self-determination and preference for challenge. Unpublished doctoral dissertation, University of Rochester, Rochester, NY.
340
REFERENCES
Hallett, G.L. (1991). Essentialism: A Wittgensteinian critique. Albany, NY: SUNY Press. Hamilton, D.L. (1981). Illusory correlation as a basis for stereotyping. In D.L.Hamilton (Ed.) Cognitive processes in stereotyping and intergroup behavior (pp. 115–144). Hillsdale, NJ: Erlbaum. Hamilton, D.L., Dugan, P.M., & Trolier, T. K. (1985). The formation of stereotypic beliefs: Further evidence for distinctiveness-based illusory correlations. Journal of Personality and Social Psychology, 48, 5–17. Hamilton, D.L., & Gifford, R.K. (1976). Illusory correlation in intergroup perception: A cognitive basis of stereotypic judgments. Journal of Experimental Social Psychology, 12, 392–407. Hamilton, D.L., & Sherman, J.W. (1994). Stereotypes. In R.S.Wyer, Jr., & T.K.Srull (Eds.), Handbook of social cognition (2nd ed., Vol. 2, pp. 1–68). San Diego, CA: Academic Press. Hamilton, D.L., & Sherman, S.J. (1996). Perceiving persons and groups. Psychological Review, 103, 336–355. Hamilton, D.L., Sherman, S.J., & Castelli, L. (2002). A group by any other name…: The role of entitativity in group perception. In W.Stroebe & M.Hewstone (Eds.), European review of social psychology (Vol. 12, pp. 139–166). Chichester, UK: Wiley. Hamilton, D.L., Sherman, S.J., & Lickel, B. (1998). Perceiving social groups: The importance of the entitativity continuum. In C. Sedikides, J.Schopler & C.A.Insko (Eds.) Intergroup cognition and intergroup behaviour (pp. 47–74). Mahwah, NJ: Erlbaum. Hamilton, D.L., Sherman, S.J., & Maddox, K.B. (1999). Dualities and continua: Implications for understanding perceptions of persons and groups. In S.Chaiken & Y. Trope (Eds.), Dual process theories in social psychology (pp. 606–626). New York: Guilford. Hamilton, D.L., & Trolier, T.K. (1986). Stereotypes and stereotyping: An overview of the cognitive approach. In J.F.Dovidio & S.L.Gaertner (Eds.), Prejudice, discrimination, and racism. New York: Academic Press. Harasty, A.S. (1996). Perceiving groups as entities: The role of “entitativity” for impression formation processes and stereotype use. Unpublished doctoral dissertation, Ohio State University. Hardin, C.D., & Higgins, E.T. (1996). Shared reality: How social verification makes the subjective objective. R.M.Sorrentino, & E. T.Higgins (Eds.), Handbook of motivation and cognition (Vol. 3, pp. 28–84). New York: Guilford Press. Harkins, S.G., Harvey, J.H., Keithly, L., & Rich, M. (1977). Cognitive tuning, encoding, and the attribution of causality. Memory and Cognition, 5, 561–565. Harmon-Jones, E., Greenberg, J., Solomon, S., & Simon L. (1996). The effects of mortality salience on intergroup bias between minimal groups. European Journal of Social Psychology, 26, 677–681. Harris, C. (1997, November 24). More whites filing for discrimination. Federal Times, pp. 1–2. Harvey, J.H., Harkins, S.G., & Kagehiro, D. K. (1976). Cognitive tuning and the attribution of causality. Journal of Personality and Social Psychology, 34, 708–715. Hashtroudi, S., Mutter, S.A., Cole, E.A., & Green, S.K. (1984). Schema-consistent and schema-inconsistent information: Processing demands. Personality and Social Psychology Bulletin, 10, 269–278. Haslam, N. (1998a). Categories of social relationship. Cognition, 53, 59–90. Haslam, N. (1998b). Natural kinds, human kinds, and essentialism. Social Research, 65, 291–314. Haslam, N. (2000). Psychiatric categories as natural kinds: Essentialist thinking about mental disorder. Social Research, 67, 1031– 1058. Haslam, N. (2002). Essentialist thinking about depression: Evidence for polarized beliefs. Manuscript submitted for publication. Haslam, N. (in press). Kinds of kinds: A conceptual taxonomy of psychiatric categories. Philosophy, Psychiatry, and Psychology. Haslam, N., & Ernst, D. (2002). Essentialist beliefs about mental disorders. Journal of Social and Clinical Psychology, 21, 682–711. Haslam, N., Rothschild, L., & Ernst, D. (2000). Essentialist beliefs about social categories. British Journal of Social Psychology, 39, 127–139. Haslam, N., Rothschild, L., & Ernst, D. (2002). Are essentialist beliefs associated with prejudice? British Journal of Social Psychology, 41, 87–100.
REFERENCES
341
Haslam, S.A. (1997). Stereotyping and social influence: Foundations of stereotype consensus. In R.Spears, P.J.Oakes, N.Ellemers & S.A.Haslam (Eds.) The social psychology of stereotyping and group life (pp. 119– 143). Oxford: Blackwell. Haslam, S.A. (2001). Psychology in organizations: The social identity approach. London: Sage. Haslam, S.A., McGarty, C., & Brown, P.M. (1996). The search for differentiated meaning is a precursor to illusory correlation. Personality and Social Psychology Bulletin, 22, 611–619. Haslam, S.A., Oakes, P.J., McGarty, C.Turner, J.C., & Onorato, R. (1995). Contextual changes in the prototypicality of extreme and moderate outgroup members. European Journal of Social Psychology, 25, 509–530. Haslam, S.A., Oakes, P.J., McGarty, C., Turner, J.C., Reynolds, K.J., & Eggins, R. (1996). Stereotyping and social influence: The mediation of stereotype applicability and sharedness by the views of ingroups and outgroup members. British Journal of Social Psychology, 35, 369–97. Haslam, S.A., Oakes, P.J., Turner, J.C., & McGarty, C. (1995). Social categorization and group homogeneity: Changes in the perceived applicability of stereotype content as a function of comparative context and trait favourableness. British Journal of Social Psychology, 34, 139–160. Haslam, S.A., Oakes, P.J., Turner, J.C., & McGarty, C. (1996). Social identity, self-categorization, and the perceived homogeneity of ingroups and outgroups. Interaction between social motivation and cognition. In R.M.Sorrentino, & T.Higgins (Eds.), Handbook of motivation and cognition (pp. 182–221). New York: Guilford Press. Haslam, S.A., Oakes, P.J., Turner, J.C., McGarty, C., & Reynolds, K.J. (1998). The group as a basis for emergent stereotype consensus. In W.Stroebe & M.Hewstone (Eds.) The European Review of Social Psychology (Vol. 8, pp. 203–239). Chichester, UK: Wiley. Haslam, S.A., & Turner, J.C. (1992) Context dependent variation in social stereotyping 2: The relationship between frame of reference, self-categorization and accentuation. European Journal of Social Psychology, 22, 251–277. Haslam, S.A., Turner, J.C., Oakes, P.J., Reynolds, K.J., & Doosje, B.J. (2002). From personal pictures in the head to collective tools in the world: How shared stereotypes allow groups to represent and change social reality. In C.McGarty, V.Y.Yzerbyt & R. Spears (Eds.) Stereotypes as explanations: The formation of meaningful beliefs about social groups (pp. 157–185). Cambridge: Cambridge University Press. Haslam, S.A., Turner, J.C., Oakes, P.J., Reynolds, K.J., Eggins, R.A., Nolan, M., et al. (1998). When do stereotypes become really consensual? Investigating the group-based dynamics of the consensualization process. European Journal of Social Psychology, 28, 755–776. Hastie, R., & Kumar, P.A. (1979). Person memory: Personality traits as organizing principles in memory for behaviors. Journal of Personality and Social Psychology, 37, 25– 38. Hastie, R., & Park, B. (1986). The relationship between memory and judgment depends on whether the judgment task is memory-based or on-line. Psychological Review, 93, 258– 268. Hecht, M., & LaFrance, M. (1998). License or obligation to smile: The effect of power and sex on amount and type of smiling. Personality and Social Psychology Bulletin, 24, 1332–1342. Heider, F. (1958). The psychology of interpersonal relations. New York: Wiley. Herek, G.M. (1988). Heterosexuals’ attitudes toward lesbians and gay men: Correlates and gender differences. Journal of Sex Research, 25, 451–457. Hewstone, M. (1990). The “ultimate attribution error”? A review of the literature on intergroup causal attribution. European Journal of Social Psychology, 20, 311–335. Hewstone, M. (1994). Revision and change of stereotypic beliefs: In search of the elusive subtyping model. In W.Stroebe & M. Hewstone (Eds.), European review of social psychology (Vol. 5, pp. 69–109). Chichester, UK: Wiley. Hewstone, M., & Crisp, R.J. (2000). Perceptions of gender group variability in majority and minority contexts: Two field studies with nurses and police officers. Manuscript in preparation. Hewstone, M., Crisp, R.J., Contarello, A., Conway, L., Voci, A., Marletta, G., et al. (2000). Perceptual processes and interaction dynamics in settings with “skewed,” “tilted,” and “balanced” sex ratios: Two field studies of women and men in academia. Manuscript in preparation.
342
REFERENCES
Hewstone, M., Crisp, R.J., Richards, Z., Voci, A., & Rubin, M. (2000). Gender, group size, and perceived group variability. Manuscript in preparation. Hewstone, M., Islam, M.R., & Judd, C.M. (1993). Models of crossed categorization and intergroup relations. Journal of Personality and Social Psychology, 64, 779–793. Hewstone, M., Johnston, L., & Aird, P. (1992). Cognitive models of stereotype change: Perceptions of homogeneous and heterogeneous groups. European Journal of Social Psychology, 22(3), 235–249. Hewstone, M., & Lord, C.G. (Eds.) (1998). Changing intergroup cognitions and intergroup behavior: The role of typicality. Mahwah, NJ: Erlbaum. Hewstone, M., Macrae, C.N., Griffiths, R., Milne, A.B., & Brown, R. (1994). Cognitive models of stereotype change: (5) Measurement, development and consequences of subtyping. Journal of Experimental Social Psychology, 30, 505–526. Hewstone, M., Martin, R., Hammer-Hewstone, C., Crisp, R.J., & Voci, A. (2001). Majority-minority relations in organizations: Challenges and opportunities. In M.A. Hogg & D.J.Terry (Eds.), Social identity processes in organizational contexts (pp. 67– 86). Philadelphia: Psychology Press. Higgins, E.T. (1987). Self-discrepancy: A theory relating self and affect. Psychological Review, 94, 319–340. Higgins, E.T. (1992). Achieving “shared reality” in the communication game: A social action that creates meaning. Journal of Language and Social Psychology, 11, 107–125. Higgins, E.T. (1996). Knowledge activation: accessibility, applicability, and salience. In E.T.Higgins & A.W.Kruglanski (Eds.), Social psychology. Handbook of basic principles (pp. 133–168). New York: Guilford Press. Higgins, E.T., McCann, C.D., & Fondacaro, R. (1982). The “communication game”: Goal-directed encoding and cognitive consequences. Social Cognition, 1, 21–37. Higgins, E.T., & Rholes, W.S. (1978). “Saying is believing”: Effects of message modification on memory and liking for the person described. Journal of Experimental Social Psychology, 14, 363–378. Higgins, R.L., & Snyder, C.R. (1989). The business of excuses. In R.Giacalone & P. Rosenfeld (Eds.), Impression management in the organization (pp. 73–85). Hillsdale, NJ: Erlbaum. Hill, C.T., Rubin, Z., & Peplau, L.A. (1976). Breakups before marriage: The end of 103 affairs. Journal of Social Issues, 32, 147–168. Hilton, J.L., Klein, J.G., & von Hippel, W. (1991). Attention allocation and impression formation. Personality and Social Psychology Bulletin, 17, 548–559. Hirschfeld, L.A. (1995). Do children have a theory of race? Cognition, 54, 209–252. Hirschfeld, L.A. (1996). Race in the making: Cognition culture and the child’s construction of the human kind. Cambridge, MA: MIT Press. Hirschfeld, L.A. (1999). La règle de la goutte de sang ou comment l’idée de race vient aux enfants. L’homme, 150, 11–40. Hoffman, C., & Hurst, N. (1990). Gender stereotypes: Perception or rationalization? Journal of Personality and Social Psychology, 58, 197–208. Hoffman, L., Clinebell, S., & Kilpatrick, J. (1997). Office romances: The new battle ground over employees’ rights to privacy and the employer’s right to intervene. Employee Responsibilities and Rights Journal, 10, 263– 275. Hofstede, G. (1980). Culture’s consequences. Thousand Oaks, CA: Sage. Hogg, M.A. (1993). Group cohesiveness: A critical review and some new directions. In W.Stroebe & M.Hewstone (Eds.), European review of social psychology (Vol. 4, pp. 85–111). Chichester, UK: Wiley. Hogg, M.A. (2000a). Subjective uncertainty-reduction through self-categorization: A motivational theory of social identity processes and group phenomena. In W.Stroebe & M.Hewstone (Eds.), European review of social psychology (Vol. 11, pp. 223–255). Chichester, UK: Wiley. Hogg, M.A. (2000b). Social identity and social comparison. In J.Suls & L.Wheeler (Eds.), Handbook of social comparison: Theory and research (pp. 401–421). New York: Kluwer/Plenum. Hogg, M.A. (2001a). Self-categorization and subjective uncertainty resolution: Cognitive and motivational facets of social identity and group membership. In J.P.Forgas, K.D. Williams, & L.Wheeler (Eds.), The social mind: Cognitive and motivational aspects of interpersonal behavior (pp. 323–349). New York: Cambridge University Press.
REFERENCES
343
Hogg, M.A. (2001b). Social categorization, depersonalization, and group behavior. In M.A.Hogg & R.S.Tindale (Eds.), Blackwell handbook of social psychology: Group processes (pp. 56–85). Oxford: Blackwell. Hogg, M.A. (2001c). Social identity and the sovereignty of the group: A psychology of belonging. In C.Sedikides & M.B.Brewer (Eds.), Individual self, relational self, collective self (pp. 123–143). Philadelphia: Psychology Press. Hogg, M.A. (2001d). A social identity theory of leadership. Personality and Social Psychology Review, 5, 184–200. Hogg, M.A. (2001e). From prototypicality to power: A social identity analysis of leadership. In S.R.Thye, E.J.Lawler, M.W.Macy, & H.A.Walker (Eds.), Advances in group processes (Vol. 18, pp. 1–30). Oxford: Elsevier. Hogg, M.A. (2003). Social identity. In M.R. Leary & J.P.Tangney (Eds.), Handbook of self and identity (pp. 462–479). New York: Guilford Press. Hogg, M.A., & Abrams, D. (1988). Social identifications: A social psychology of intergroup relations and group processes. London: Routledge. Hogg, M.A., & Abrams, D. (1990). Social motivation, self-esteem, and social identity. In D.Abrams & M.A.Hogg (Eds.), Social identity theory: Constructive and critical advances (pp. 28–47). London: Harvester Wheatsheaf, & New York: Springer-Verlag. Hogg, M.A., & Abrams, D. (1993). Towards a single-process uncertainty-reduction model of social motivation in groups. In M.A.Hogg & D.Abrams, (Eds.), Group motivation: Social psychological perspectives (pp. 173– 190). Hemel Hempstead, UK: Harvester Wheatsheaf, and New York: Prentice Hall. Hogg, M.A., & Abrams, D. (1999). Social identity and social cognition: Historical background and current trends. In D. Abrams & M.A.Hogg (Eds.), Social identity and social cognition (pp. 1–25). Oxford: Blackwell. Hogg, M.A., Fielding, K.S., & Darley, J. (in press). Fringe dwellers: Processes of deviance and marginalization in groups. In D. Abrams, J.Marques, & M.A.Hogg (Eds.), Social psychology of inclusion and exclusion. Philadelphia: Psychology Press. Hogg, M.A., & Grieve, P. (1999). Social identity theory and the crisis of confidence in social psychology: A commentary and some research on uncertainty reduction. Asian Journal of Social Psychology, 2, 79–93. Hogg, M.A., & Hains, C.S. (1998). Friendship and group identification: A new look at the role of cohesiveness in group think. European Journal of Social Psychology, 28, 323–341. Hogg, M.A., & Hornsey, M.J. (1998). Pauline Hanson’s One Nation: A social psychological analysis. Unpublished manuscript, University of Queensland, Australia. Hogg, M.A., & McGarty, C. (1990). Self-categorization and social identity. In D.Abrams & M.A.Hogg (Eds.) Social identity theory: Constructive and critical advances (pp. 10– 27). London and New York: Harvester Wheatsheaf and Springer-Verlag. Hogg, M.A., & Mullin, B.A. (1999). Joining groups to reduce uncertainty: Subjective uncertainty reduction and group identification. In D.Abrams & M.A.Hogg (Eds). Social identity and social cognition (pp. 249– 279). Oxford: Blackwell. Hogg, M.A., & Mullin, B.A. (2000). Reducing subjective uncertainty by group identification: The role of group relevance. Unpublished manuscript, University of Queensland, Australia. Hogg, M.A., & Reid, S. (1998). Uncertainty, social identification, and extremism in Australia. Unpublished manuscript, University of Queensland, Australia. Hogg, M.A., & Reid, S.A. (2001). Social identity, leadership, and power. In A.Y.Lee-Chai & J.A.Bargh (Eds.), The use and abuse of power: Multiple perspectives on the causes of corruption (pp. 159–180). Philadelphia: Psychology Press. Hogg, M.A., & Sussman, K. (1999). Uncertainty, entitativity, and group identification: Studies of student affiliation with campus clubs and societies. Unpublished manuscript, University of Queensland, Australia, and Princeton University. Hogg, M.A., & van Knippenberg, D. (2003). Social identity and leadership processes in groups. In M.P.Zanna (Ed.), Advances in experimental social psychology (Vol. 35, pp. 1–92)). San Diego, CA: Academic Press. Hogg, M.A., & Williams, K. (2000). From I to we: Social identity and the collective self. Group Dynamics: Theory, Research, and Practice, 4, 81–97.
344
REFERENCES
Holtz, R., & Miller, N. (1985). Assumed similarity and opinion certainty. Journal of Personality and Social Psychology, 48, 890–898. Hong, Y. (1994). Predicting trait versus process inferences: The role of lay theories. Unpublished doctoral dissertation, Columbia University. Hong, Y., Chiu, C., Dweck, C.S., & Sacks, R. (1997). Lay theories and evaluative processes in person cognition, Journal of Experimental Social Psychology, 33, 296–323. Hong, Y., Chiu, C., Dweck, C.S., Lin, D.M. S., & Want, W. (1999). Implicit theories attributions, and coping: A meaning systems approach. Journal of Personality and Social Psychology, 77, 588–599. Hong, Y., Chiu, C., Yeung, G., & Tong, Y. (1999). Social comparison during political transition: Interaction of entity versus incremental beliefs and social identities. International Journal of Intercultural Relations, 23, 257–280. Hong, Y., Levy, S.R., & Chiu, C. (2001). The contribution of the lay theories approach to the study of groups. Personality and Social Psychology Review, 5, 98–106. Hong, Y., & Wong, R. (2000). The individual versus group agency measure. Unpublished new data, Hong Kong University of Science and Technology. Hoover, N.C. (1999). National survey: Initiation rites and athletics for NCAA sports teams. Alfred, NY: Alfred University. Hornsey, M.J., & Hogg, M.A. (2000). Assimilation and diversity: An integrative model of subgroup relations. Personality and Social Psychology Review, 4, 143–156. Hovland, C.I., Janis, I.L., & Kelley, H.H. (1953). Communication and persuasion. London: Oxford University Press. Hoyle, R.H., Pinkley, R.L., & Insko, C.A. (1989). Perceptions of behavior: Evidence for differing expectations for interpersonal and intergroup interactions. Personality and Social Psychology Bulletin, 15, 365–376. Hull, D.L. (1965) The effect of essentialism on taxonomy: Two thousand years of stasis. The British Journal for the Philosophy of Science, 15, 314–326. Hutchison, P. (2000). The effects of pro-norm and anti-norm deviance on judgements of groups and their members. Unpublished MSc. thesis, University of Kent at Canterbury. Hutchison, P., & Abrams, D. (in press). Ingroup identification moderates stereotype change in reaction to in-group deviance. European Journal of Social Psychology. Hutchison, P., & Abrams, D., & Viki, G.T. (2002). Social identification, social attraction, and reactions to deviant ingroup members. Paper presented at the Society of Australasian Social Psychologists 8th Annual Meeting, Adelaide, Australia, 25–28 April. Insko, C.A., Drenan, S., Solomon, M.R., Smith, R., & Wade, T. (1983). Conformity as a function of the consistency of positive self-evaluation with being liked and being right. Journal of Experimental Social Psychology, 19, 341–358. Insko, C.A., Hoyle, R.H., Pinkley, R.L., Hong, G., Slim, R.M., Dalton, B., et al. (1988). Individual-group discontinuity: The role of a consensus rule. Journal of Experimental Social Psychology, 24, 505–519. Insko, C.A., Pinkley, R.L., Hoyle, R.H., Dalton, B., Hong, G., Slim, R.M., et al. (1987). Individual versus group discontinuity: The role of intergroup contact. Journal of Experimental Social Psychology, 23, 250– 267. Insko, C.A., & Schopler, J. (1998). Differential distrust of groups and individuals. In C. Sedikides, J.Schopler, & C.A.Insko (Eds.), Intergroup cognition and intergroup behavior (pp. 75–108). Mahwah, NJ: Erlbaum. Insko, C.A., Schopler, J., Drigotas, S.M., Graetz, K.A., Kennedy, J., Cox, C., et al. (1993). The role of communication in interindividual-intergroup discontinuity. Journal of Conflict Resolution, 37, 108–138. Insko, C.A., Schopler, J., Gaertner, L., Wildschut, T, Kozar, R., Pinter, B., et al. (2001). Interindividual-intergroup discontinuity reduction through the anticipation of future interaction. Journal of Personality and Social Psychology, 80, 95–111. Insko, C.A., Schopler, J, Graetz, K.A., Drigotas, S.M., Currey, D.P., Smith, S.L., et al. (1994). Interindividualintergroup discontinuity in the Prisoner s Dilemma Game. Journal of Conflict Resolution, 38, 87–116. Insko, C.A., Schopler, J., Hoyle, R.H., Dardis, G.J., & Graetz, K.A. (1990). Individualgroup discontinuity as a function of fear and greed. Journal of Personality and Social Psychology, 58, 68–79.
REFERENCES
345
Insko, C.A., Schopler, J., Kennedy, J.F., Dahl, K.R., Graetz, K.A., & Drigotas, S.M. (1992). Individual-group discontinuity from the differing perspectives of Campbell’s realistic group conflict theory and Tajfel and Turner’s social identity theory. Social Psychology Quarterly, 55, 272–291. Insko, C.A., Schopler J., Pemberton M.B., Wieselquist, J., McIlraith S.A., Currey, D. P., et al. (1998). Future consequences and the reduction of interindividual-intergroup discontinuity. Journal of Personality and Social Psychology, 75, 695–711. Insko, C.A., Schopler, J., & Sedikides, C. (1998). Personal control, entitativity, and evolution. In C.Sedikides, J.Schopler, & C. A.Insko (Eds.), Intergroup cognition and intergroup behavior (pp. 110–120). Mahwah, NJ: Erlbaum. Insko, C.A., Wildschut, T., & Pinter, B. (2003). Interindividual-intergroup discontinuity as a function of ingroup and outgroup procedural interdependence. Manuscript submitted for publication. Irving, P.G., & Meyer, J.P. (1994). Reexamination of the met-expectations hypothesis: A longitudinal analysis. Journal of Applied Psychology, 79, 937–949. Islam, M.R., & Hewstone, M. (1993a). Dimensions of contact as predictors of intergroup anxiety, perceived out-group variability, and out-group attitude: An integrative model. Personality and Social Psychology Bulletin, 19, 700–710. Islam, M.R., & Hewstone, M. (1993b). Intergroup attributions and affective consequences in majority and minority groups. Journal of Personality and Social Psychology, 64, 936–950. Jackson, L.A., & Hymes, R.W. (1985). Gender and social categorization: Familiarity and ingroup polarization in recall and evaluation. Journal of Social Psychology, 125, 81–88. James, W. (1890). The principles of psychology. New York: Holt, Rinehart & Winston. Janis, I.L. (1972). Victims of groupthink: A psychological study of foreign-policy decisions and fiascos. Boston: Houghton Mifflin. Janis, I.L. (1982). Groupthink (2nd ed.). Boston: Houghton Mifflin. Jarvis, G., & Petty, R.E. (1996). The need to evaluate. Journal of Personality and Social Psychology, 70, 172–196. Jetten, J., Branscombe, N.R., Schmitt, M.T., & Spears, R. (2001). Rebels with a cause: Group identification as a response to perceived discrimination from the mainstream. Personality and Social Psychology Bulletin, 27, 1204–1213. Jetten, J., Hogg, M.A., & Mullin, B.-A. (2000). Ingroup variability and motivation to reduce subjective uncertainty. Group Dynamics: Theory, Research, and Practice, 4, 184–198. Jetten, J., & Spears, R. (in press). The divisive potential of differences and similarities: The role of intergroup distinctiveness in intergroup differentiation. In W.Stroebe & M. Hewstone (Eds.), European Review of Social Psychology. Chichester, UK: Wiley. Jetten, J., Spears, R., & Manstead, A.S. R. (1997). Ingroup norms and intergroup discrimination distinctive selfcategorization and social identity effects. Journal of Personality and Social Psychology, 71, 1222– 1233. Jetten, J., Spears, R., & Manstead, A.S.R. (1998). Defining dimensions of distinctiveness: Group variability makes a difference to differentiation. Journal of Personality and Social Psychology, 74, 1481–1492. Jetten, J., Spears, R., & Manstead, A.S.R. (2001). Similarity as a source of discrimination: The role of group identification. European Journal of Social Psychology, 31, 621– 640. Johnson, C., & Mullen, B. (1993). The determinants of differential group evaluations in distinctiveness-based illusory correlations in stereotyping. British Journal of Social Psychology, 32, 253–263. Johnson, J.A., Gerner, C.K., Efran, J.S., & Overton, W.F. (1988). Personality as a basis for theoretical predilections. Journal of Personality and Social Psychology, 55, 824–835. Johnson, M.P., & Milardo, R.M. (1984). Network interference in pair relationships: A social recasting of Slater’s (1963) theory of social regression. Journal of Marriage and the Family, 46, 893–899. Jones, E.E., & Harris, V.A. (1967). The attribution of attitudes. Journal of Experimental Social Psychology, 3, 1–24. Jones, E.E., Wood, G.C., & Quattrone, G.A. (1981). Perceived variability of personal characteristics in in-groups and out-groups: The role of knowledge and evaluation. Personality and Social Psychology Bulletin, 7, 523–528.
346
REFERENCES
Jones, G.E. (1999). Hierarchical workplace romance: An experimental examination of team member perceptions. Journal of Organizational Behavior, 20, 1059–1072. Jones, J.M. (1972). Prejudice and racism. Reading, Mass: Addison-Wesley. Jost, J.T. (1995). Negative illusions: Conceptual clarification and psychological evidence concerning false consciousness. Political Psychology, 16, 397–424. Jost, J.T., & Banaji, M.R. (1994). The role of stereotyping in system-justification and the production of falseconsciousness. British Journal of Social Psychology, 33, 1–27. Judd. C.M., & Brauer, M. (1995). Repetition and evaluative extremity. In R.E.Petty & J. A.Krosnick (Eds.) Attitude strength: Antecedents and consequences. Hillsdale, NJ: Erlbaum. Judd, C.M., & Lusk, C.M. (1984). Knowledge structures and evaluative judgments: Effects of structural variables on judgmental extremity. Journal of Personality and Social Psychology, 46, 1193–1207. Judd, C.M., & Park, B. (1988). Out-group homogeneity: Judgments of variability at the individual and group levels. Journal of Personality and Social Psychology, 54, 778–788. Judd, C.M., Park, B., Ryan, C.S., Brauer, M., & Kraus, S. (1995). Stereotypes and ethnocentrism: Diverging interethnic perceptions of African American and White American youth. Journal of Personality and Social Psychology, 69, 460–481. Judd, C.M., Ryan, C.S., & Park, B. (1991). Accuracy in the judgment of ingroup and ougroup variability. Journal of Personality and Social Psychology, 61, 366–379. Kahneman, D., & Miller, D.T. (1986). Norm theory: Comparing reality to its alternatives. Psychological Review, 93, 136–153. Kanter, R.M. (1977). Some effects of proportions on group life: Skewed sex ratios and responses to token women. American Journal of Sociology, 82, 465–490. Kaplan, S. (1978). Concerning the power of content-identifying methodologies. In E.H. Zube, & T.Daniel (Eds.), Environmental aesthetics. Washington, DC: U.S. Forest Service. Karenfentis, D.M., & Levy, S.R. (2003). The role of children’s lay theories about the malleability of human attributes in beliefs about volunteering for disadvantaged groups. Manuscript submitted for publication. Kashima, Y. (2000a). Conceptions of culture and person for psychology. Journal of Cross-Cultural Psychology, 31, 14–31. Kashima, Y. (2000b). Maintaining cultural stereotypes in the serial reproduction of narratives. Personality and Social Psychology Bulletin, 26, 594–604. Kashima, Y. (2001). Culture and social cogni tion: Towards a social psychology of cultural dynamics. In D.Matsumoto (Ed.), Handbook of culture and psychology (pp. 325– 360). New York: Oxford University Press. Kashima, Y., Kashima, E., Chiu, C.-Y., Farsides, T., Gelfand, M., Hong, Y.-Y., et al. (2001). Culture entitativitive, and agency: Are individuals universally believed to be more real entities than groups? Unpublished manuscript, Department of Psychology, the University of Melbourne, Australia. Kashima, Y., & Vears, C. (2000). Communication and entitativity: Establishing social reality of a group by communication. Unpublished manuscript, Department of Psychology, the University of Melbourne, Australia. Katz, D., & Braly, K. (1933). Racial stereotypes of one hundred college students. Journal of Abnormal and Social Psychology, 28, 280– 290. Katz, I., & Hass, R.G. (1988). Racial ambivalence and American value conflict: Correlational and priming studies of dual cognitive structures. Journal of Personality and Social Psychology, 55, 893–905. Kaul, T.J., & Bednar, R.L. (1994). Pretraining and structure: Parallel lines yet to meet. In A.Fuhriman & G.Burlingame (Eds.), Handbook of group psychotherapy: An empirical and clinical synthesis (pp. 155–188). New York: Wiley. Kayser, K. (1993). When love dies: The process of marital dissatisfaction. New York: Guilford Press. Keen, S. (1986). Faces of the enemy. Reflections of the hostile imagination. San Francisco, CA: Harper & Row. Keil, F.C. (1989). Concepts, kinds, and cognitive development. Cambridge, MA: MIT Press. Kelley, H.H. (1967). Attribution theory in social psychology. In D.Levine (Ed.), Nebraska symposium on motivation (Vol. 15, pp. 192–238). Lincoln: University of Nebraska Press. Kelley, H.H., & Thibaut, J.W. (1978). Interpersonal relations. New York: Wiley.
REFERENCES
347
Kelly, C. (1988). Intergroup differentiation in a political context. British Journal of Social Psychology, 27, 319–332. Kelly, C. (1989). Political identity and perceived intragroup homogeneity. British Journal of Social Psychology, 28, 239–250. Kelly, C. (1993) Group identification, intergroup perceptions, and collective action. In W.Stroebe & M.Hewstone Eds.), European review of social Psychology (Vol. 4, pp. 59– 83). Chichester, UK: Wiley. Kelly, G.A. (1955). The psychology of personal constructs. New York: Norton. Kelly, J.B. (2000). Children’s adjustment in conflicted marriage and divorce: A decade review of research. Journal of the American Academy of Child and Adolescent Psychiatry, 39, 963–973. Kelman, H.C. (1969). Patterns of personal involvement n the national system: A social psychological analysis of political legitimacy. In J.N.Rosenau (Ed.), International politics and foreign policy. A reader in research and theory (pp. 276–288). Glencoe, IL: Free Press. Keltner, D., Young, R.C., Oemig, C., Heerey, E.A., & Monarch, N. (1998). Teasing in hierarchical and intimate relations. Journal of Personality and Social Psychology, 75, 1231– 1247. Kenny, D.A. (1994). Interpersonal perception: A social relations analysis. New York: Guilford Press. Kerr, N.L. (1999). Anonymity and social control in social dilemmas. In M.Foddy, M. Smithson, S.Schneider, & M.Hogg (Eds.), Resolving social dilemmas: Dynamic, structural, and intergroup aspects (pp. 103–119). Philadelphia: Psychology Press. Khalaf, A., & McGarty, C. (2002). Understanding and changing mental illness stereotypes I: The group-type distinction. Unpublished manuscript, Australian National University. Kitzinger, C. (1991). Lesbians and gay men in the workplace: Psychosexual issues. In M.J. Davidson & J.Earnshaw (Eds.), Vulnerable workers: Psychosocial and legal issues (pp. 223–240). New York: Wiley. Klauer, K.C., Ehrenberg, K., & Wegener, I. (2003). Crossed categorization and stereotyping: Structural analyses effect patterns, and dissociative effects of context relevance. Journal of Experimental Social Psychology, 39, 332–354. Klauer, K.C., & Meiser, T. (2000). A source-monitoring analysis of illusory correlation. Personality and Social Psychology Bulletin, 26, 1074–1093. Klauer, K.C., & Wegener, I. (1998). Unraveling social categorization in the “Who said what?” paradigm. Journal of Personality and Social Psychology, 75, 1155–1178. Klauer, K.C., Wegener, I., & Ehrenberg, K. (2002). Perceiving minority members as individuals: The effects of relative group size in social categorization. European Journal of Social Psychology, 32, 223–245. Klein, O., Azzi, A.E., Brito, R., & Berckmans, S. (1999, June). Nationalism and the strategic presentation of identity. Paper presented at the meeting of the Akademie van Wetenshappen, Amsterdam. Klein, S.B., & Loftus, J. (1993). The mental representation of trait and autobiographical knowledge about the self. T.K.Srull & R.S. Wyer (Eds.), Advances in social cognition (Vol. 5, pp. 1–49). Hillsdale, NJ: Erlbaum. Kipnis, D. (1972). Does power corrupt? Journal of Personality and Social Psychology, 24, 33–41. Koestner, R., Ryan, R.M., Bernieri, F., & Holt, K. (1984). Setting limits in children’s behavior: The differential effects of controlling versus informational styles on intrisic motivation and creativity. Journal of Personality, 52, 233–248. Kofta, M., & Sedek, G. (1999). Group entitativity, agency schema, and group soul stereotypes. Paper presented at EAESP conference on entitativity, Louvain-la-Neuve, Belgium. Kornblith, H. (1993). Inductive inference and its natural ground: An essay in naturalistic epistemology. Cambridge, MA: MIT Press. Kramer, D., Ber, R., & Moore, M. (1987). Impact of workshop on students’ and physicians’ rejecting behaviors in patient interviews. Journal of Medical Education, 62, 904–910. Kraus, S., Ryan, C.S., Judd, C.M., Hastie, R., & Park, B. (1993). Use of mental frequency distributions to represent variability among members of social categories. Social Cognition, 11, 22–43. Kretch, D., & Crutchfield, R.C. (1948). Theory and problems of social psychology. New York: McGraw-Hill. Kripke, S. (1980). Naming and necessity. Cambridge, MA: Harvard University Press. Krueger, J. (1998). On the perception of social consensus. Advances in Experimental Social Psychology, 30, 163–240.
348
REFERENCES
Krueger, J., & Clement, R.W. (1994). Memory-based judgments about multiple categories: A revision and extension of Tajfel’s accentuation theory. Journal of Personality and Social Psychology, 67, 35–47. Krueger, J., & Clement, R.W. (1996). Inferring category characteristics from sample characteristics: Inductive reasoning and social projection. Journal of Experimental Psychology: General, 125, 52–68. Krueger, J., & Rothbart, M. (1988). Use of categorical and individuating information in making inferences about personality. Journal of Personality and Social Psychology, 55, 187–195. Kruglanski, A.W. (1989). Lay epistemics and human knowledge: Cognitive and motivational bases. New York: Plenum Press. Kruglanski, A.W., & Webster, D.M. (1996). Motivated closing of the mind: “Seizing” and “freezing.” Psychological Review, 103, 263– 283. Kunda, Z., & Oleson, K.C. (1995). Maintaining stereotypes in the face of disconfirmation: Constructing grounds for subtyping deviants. Journal of Personality and Social Psychology, 68, 565–579. Kunda, Z., & Oleson, K.C. (1997). When exceptions prove the rule: How extremity of deviance determines the impact of deviant examples on stereotypes. Journal of Personality and Social Psychology, 72, 965–979. Lakoff, G. (1987). Women, fire, and dangerous things: What categories reveal about the human mind. Chicago: University of Chicago Press. Lalonde, R.N., & Silverman, R.A. (1994). Behavioral preferences in response to social injustice: The effects of group permeability and social identity salience. Journal of Personality and Social Psychology, 66, 78– 85. Lambert, A.J. (1995). Stereotypes and social judgment: The consequences of group variability. Journal of Personality and Social Psychology, 68, 388–403. Lambert, A.J., Barton, L.R., Lickel, B., & Wells, J. (1998). The influence of group variability and processing goals on the ease of making judgments about social categories. Personality and Social Psychology Bulletin, 24, 807–820. Lambert, S., Hogg, M.A., Klein, O., & Azzi, A.E. (2002). Perceptual accentuation and uncertainty reduction. Manuscript submitted for publication, Université Libre de Bruxelles, Belgium. Latané, B. (1981). The psychology of social impact. American Psychologist, 36, 343–356. Latané, B., & Nida, S. (1981). Ten years of research on group size and helping. Psychological Bulletin, 89, 308–324. Lea, M., Spears, R., & De Grotte, D. (2001). Knowing me, knowing you: Effects of visual anonymity on selfcategorization, stereotyping, and attraction in computer-mediated groups. Personality and Social Psychology Bulletin, 27, 526–537. Leahey, T.H. (1987). A history of psychology: Main vurrents in psychological thought. New Jersey: Prentice-Hall. Leary, M.R. (1995). Self-presentation: Impression management and interpersonal behavior. Madison, WI: Brown and Benchmark. Leary, M.R., Tambor, E.S., Terdal, S.K., & Downs, D.L. (1995). Self-esteem as an interpersonal monitor: The sociometer hypothesis. Journal of Personality and Social Psychology, 68, 518–530. Leary, M.R., & Tangney, J. (Eds.) (2002). Handbook of self and identity. New York: Guilford Press. Le Bon G., (1895). The Crowd. London: Unwin. Lee, F.S. (1994). Membership has its costs. Journal of the American Medical Association, 271, 1048–1049. Lepore, L., & Brown, R. (1997). Category and stereotype activation: Is prejudice inevitable? Journal of Personality and Social Psychology, 72, 275–287. Lester, R.E. (1987). Organizational culture, uncertainty reduction, and the socialization of new organizational members. In S.Thomas (Ed.), Culture and communication: Methodology, behavior, artifacts, and institutions (pp. 105–113). Norwood, NJ: Ablex. Leventhal, H. (1962). The effect of set and discrepancy on impression change. Journal of Personality, 30, 1–15. Levine, J.M. (1989). Reactions to opinion deviance in small groups. In P.B.Paulus (Ed.) Psychology of group influence (pp. 187– 231). Hillsdale, NJ: Erlbaum. Levine, J.M., & Kaarbo, J. (2001). Minority influence in political decision-making groups. In C.K.De Dreu & N.K.De Vries (Eds.), Group consensus and minority influence: Implications for innovation (pp. 229– 257). Oxford, UK: Blackwell.
REFERENCES
349
Levine, J.M., & Moreland, R.L. (1994). Group socialization: Theory and research. In W. Stroebe & M.Hewstone (Eds.), European Review of Social Psychology (Vol. 5, pp. 305– 336). Chichester, England: John Wiley. Levine, J.M., Moreland, R.L., & Ryan, C. (1997). Group socialization and intergroup relations. In C.Sedikides, J.Schopler, & C. Insko (Eds.), Intergroup cognition and intergroup behavior (pp. 283–308). Mahwah, NJ: Erlbaum. Levine, J.M., & Thompson, L. (1996). Conflict in groups. In E.T.Higgins & A.W. Kruglanski (Eds.) Social psychology: Handbook of basic principles (pp. 745–776). New York: The Guilford Press. Levy, S.R. (1998). Children’s static versus dynamic conceptions of people: Their impact on intergroup attitudes. Unpublished doctoral dissertation, Columbia University. Levy, S.R., & Dweck, C.S. (1998). Trait- versus process-focused social judgment. Social Cognition, 16, 151–172. Levy, S.R., & Dweck, C.S. (1999). The impact of children’s static versus dynamic conceptions of people on stereotype formation. Child Development, 70, 1163–1180. Levy, S.R., Plaks, J.E., & Dweck, C.S. (1999). Modes of social thought: Lay theories and social understanding. In S.Chaiken & Y. Trope (Eds.), Dual process theories in social psychology (pp. 179–202). New York: Guilford Press. Levy, S.R., Plaks, J.E., Hong, Y., Chiu, C., & Dweck, C.S. (2001). Static vs. dynamic theories and the perception of groups: Different routes to different destinations. Personality and Social Psychology Review, 5, 156–168. Levy, S.R., Stroessner, S.J., & Dweck, C.S. (1998). Stereotype formation and endorsement: The role of implicit theories. Journal of Personality and Social Psychology, 74, 1421–1436. Lewicki, P. (1986). Processing information about covariation that cannot be articulated. Journal of Experimental Psychology: Learning, Memory, and Cognition, 12, 135–146. Lewin, K. (1936). Principles of topological psychology. New York: McGraw-Hill. Lewin, K. (1941). Regression, retrogression and development. University of Iowa Studies of Child Welfare, 18, 1–43. Lewin, K. (1948; reprinted 1997). Resolving social conflicts: Selected papers on group dynamics. (G.Weiss Lewin (ed.)) Washington DC: American Psychological Association. Lewis, R.A. (1973). Social reaction and the formation of dyads: An interactionist approach to mate selection. Sociometry, 36, 409–418. Leyens, J.-P., & Yzerbyt, V.Y. (1992). The ingroup overexclusion effect: Impact of valence and confirmation on stereotypical information search. European Journal of Social Psychology, 22, 549–569. Leyens, J.-P, Yzerbyt, V., & Schadron, G. (1994). Stereotypes and social cognition. London: Sage. Lickel, B., Hamilton, D.L., Wieczorkowska, G., Lewis, A., Sherman, S.J., & Uhles, A. (2000). Varieties of groups and the perception of group entitativity. Journal of Personality and Social Psychology, 78, 223–246. Lindskold, S., Albert, K.D., Baer, R., & Moore, W.L. (1976). Territorial boundaries of interacting groups and passive audiences. Sociometry, 39, 71–76. Linville, P.W. (1982). The complexity-extremity effect and age-based stereotyping. Journal of Personality and Social Psychology, 42, 193–211. Linville, P.W. (1987). Self-complexity as a buffer against stress-related illness and depression. Journal of Personality and Social Psychology, 52, 663–676. Linville, P.W. (1998). The heterogeneity of homogeneity. In J.M.Darley & J.Cooper (Eds.), Attribution and social interaction: The legacy of Edward E.Jones (pp. 423– 482). Washington, DC: American Psychological Association. Linville, P.W., & Fischer, G.W. (1993). Exemplar and abstraction models of perceived group variability and stereotypicality. Social Cognition, 11, 92–125. Linville, P.W., & Fischer, G.W. (1998). Group variability and covariation: Effects on intergroup judgment and behavior. In C. Sedikides, C.A.Insko, & J.Schopler (Eds.), Intergroup cognition and intergroup behavior (pp. 123–150). Mahwah, NJ: Erlbaum. Linville, P.W., & Fischer, G.W. (2003a). Accuracy of covariation judgments. Unpublished paper, Duke University, Durham, NC. Linville, P.W., & Fischer, G.W. (2003b). Perceptions of group entitativity over time. Unpublished paper, Duke University, Durham, NC.
350
REFERENCES
Linville, P.W., Fischer, G.W., & Salovey, P. (1989). Perceived distributions of the characteristics of ingroup and outgroup members: Empirical evidence and a computer simulation. Journal of Personality and Social Psychology, 57, 165–188. Linville, P.W., Fischer, G.W., & Yoon, C. (1996). Perceived covariation among the features of ingroup and outgroup members: An outgroup covariation effect. Journal of Personality and Social Psychology, 6, 1–23. Linville, P.W., & Fischer, G.W., Zauberman, G., & Rona, K. (2003). Perceived variability, covariance, and extremity of judgments about group members. Unpublished paper, Duke University, Durham, NC. Linville, P.W., & Jones, E.E. (1980). Polarized appraisals of out-group members. Journal of Personality and Social Psychology, 38, 689–703. Linville, P.W., Salovey, P., & Fischer, G.W. (1986). Stereotyping and perceived distributions of social characteristics: An application to ingroup-outgroup perception. In J. Dovidio & S.L.Gaertner (Eds.), Prejudice, discrimination, and racism (pp.165–208). New York: Academic Press. Lippmann, W. (1922). Public opinion. New York: Harcourt Brace. Lodewijkx, H.F.M., & Syroit, J.E.M.M. (1997). Severity of initiation revisited: Does severity of initiation increase attractiveness in real groups? European Journal of Social Psychology, 27, 275–300. Lodewijkz, H.F.M., Stroebe, K., & Spears, R. (2000). Do unto others as they do unto you: Bounded reciprocation versus identification as determinants of ingroup favoritism. Manuscript submitted for publication, Unversity of Utrecht and University of Amsterdam. Long, M.K., & Spears, R. (1997). The self-esteem hypothesis revisited: Differentiation and the disaffected. In R.Spears, P.J.Oakes, N.Ellemers, & S.A.Haslam (Eds.), The social psychology of stereotyping and group life (pp. 296–317). Oxford: Blackwell. Lopes, L.L. (1987). Between hope and fear: The psychology of risk. Advances in Experimental Psychology, 20, 255–295. Lord, C.G., Desforges, D.M., Fein, S., Pugh, M.A., & Lepper, M.R. (1994). Typicality effects in attitudes toward social policies: A concept-mapping approach. Journal of Personality and Social Psychology, 66, 658–673. Lord, C.G., & Lepper, M.R. (1999). Attitude Representation Theory. In M.P.Zanna (Ed.), Advances in experimental social psychology (Vol. 31, pp. 265–343). San Diego: Academic Press. Lorenzi-Cioldi, F. (1988). Individus dominants et groupes dominés [Dominant individuals and dominated groups]. Grenoble, France: Presses Universitaires. Lorenzi-Cioldi, F. (1993). They all look alike, but so do we, sometimes… Perception of ingroup and outgroup homogeneity as a function of gender and context. British Journal of Social Psychology, 32, 111–124. Lorenzi-Cioldi, F. (1998). Group status and perceptions of homogeneity. In W. Stroebe & M.Hewstone (Eds.), European review of social psychology (Vol. 9, pp. 31–75). Chichester, UK: Wiley. Lorenzi-Cioldi, F., Deaux, K., & Dafflon, A.-C. (1998). Group homogeneity as a function of relative social status. Swiss Journal of Psychology, 57, 255–273. Lorenzi-Cioldi, F., Eagly, A.H., & Stewart, T. (1995). Homogeneity of gender groups in memory. Journal of Experimental Social Psychology, 31, 193–217. Louis, M.R. (1980). Surprise and sense making: What newcomers experience in entering unfamiliar organizational settings. Administrative Science Quarterly, 25, 226–251. Luke, D.A., Roberts, L., & Rappaport, J. (1993). Individual, group context, and individual-group fit predictors of selfhelp group attendance. Journal of Applied Behavioral Science, 29, 216–238. Lyman, S.M., & Scott, M.B. (1967). Territoriality: A neglected sociological dimension. Social Forces, 15, 236–249. Lyons, A., & Kashima, Y. (2001). The reproduction of culture: Communications tend to maintain cultural stereotypes. Social Cognition, 19, 372–394. Maass, A. (1999). Linguistic intergroup bias: Stereotype perpetuation through language. In M.Zanna (Ed.), Advances in experimental social psychology (Vol. 31, pp. 79–121). San Diego, CA: Academic Press. MacDonald, A.P. (1970). Revised scale for ambiguity tolerance: Reliability and validity. Psychological Reports, 26, 791–798.
REFERENCES
351
Mackie, D.M., Sherman, J.W., & Worth, L. T. (1993). On-line and memory-based processes in group variability judgments. Social Cognition, 11, 44–69. Macrae, C.N. & Bodenhausen, G.V. (2000). Social cognition: Thinking categorically about others. Annual Review of Psychology, 51, 93–120. Macrae, C.N., Bodenhausen, G.V., Schloerscheidt, A.M., & Milne, A.B. (1999). Tales of the unexpected: Executive function and person perception. Journal of Personality and Social Psychology, 76, 200–213. Macrae, C.N., Hewstone, M., & Griffiths, R. J. (1993). Processing load and memory for stereotype-based information. European Journal of Social Psychology, 23, 77–87. Macrae, C.N., Milne, A.B., & Bodenhausen, G.V. (1994). Stereotypes as energy-saving devices: A peek inside the cognitive toolbox. Journal of Personality and Social Psychology, 66, 37–47. Macrae, C.N., Stangor, C., & Hewstone, M. (Eds.). (1996). Stereotypes and stereotyping. New York: Guilford Press. Malle, B. (1999). How people explain behavior: A new theoretical framework. Personality and Social Psychology Review, 3, 23–48. Marger, M.N. (1985). Race and ethnic relations: American and global perspectives. Belmont, CA: Wadsworth. Marin, G. (1984). Stereotyping Hispanics: The differential effect of research method, label, and degree of contact. International Journal of Intercultural Relations, 8, 17–27. Markman, A.B. (1999). Knowledge representation. Mahwah, NJ: Mahwah. Marks, G. (1999). Territorial identities in the European Union. In Jeffrey J.Anderson (Ed.), Regional integration and democracy: Expanding on the European experience. Boulder, CO: Rowman & Littlefield. Markus, H.R., & Kitayama, S. (1991). Culture and the self: Implications for cognition, emotion, and motivation. Psychological Review, 98, 224–253. Marques, J.M. (1990). The black-sheep effect: Outgroup homogeneity in social comparison settings. In D.Abrams & M.A.Hogg (Eds.), Social identity theory: Constructive and critical advances (pp. 131–151). London: Harvester Wheatsheaf, and New York: Springer-Verlag. Marques, J.M, Abrams, D., Paez, D., & Hogg, M. (2001). Social categorization, social identification, and rejection of deviant group members. In M.A.Hogg & R.S.Tindale (Eds.), Blackwell Handbook of Social Psychology: Group Processes (pp. 400–424). Oxford: Blackwell. Marques, J.M., Abrams, D., Páez, D., & Martinez-Taboada, C. (1998). The role of categorization and ingroup norms in judgments of groups and their members. Journal of Personality and Social Psychology, 75, 976–988. Marques, J.M., Abrams, D., & Serôdio, R.G. (2001). Being better by being right: Subjective group dynamics and derogation of ingroup deviants when generic norms are undermined. Journal of Personality and Social Psychology, 81, 436–447. Marques, J.M., & Paez, D. (1994). The “black sheep effect”: Social categorization, rejection of ingroup deviates, and perception of group variability. In W.Stroebe & M. Hewstone (Eds.), European review of social psychology (Vol. 5, pp. 38–68). Chichester, UK: Wiley. Marques, J.M., Paez, D., & Abrams, D. (1998). Social identity and intragroup differentiation as subjective social control. In S.Worchel, J.F.Morales, D.Paez, & J.-C.Deschamps (Eds.), Social identity: International perspectives (pp. 124–142). London: Sage. Marques, J.M., Robalo, E.M., & Rocha, S.A. (1992). Ingroup bias and the black sheep effect: Assessing the impact of cognitive-motivational and informational antecedents of judgmental extremity towards ingroup members. European Journal of Social Psychology, 22, 331–352. Marques, J.M., & Yzerbyt, V.Y. (1988). The black sheep effect: Judgmental extremity towards ingroup members in inter- and intra-group situations. European Journal of Social Psychology, 18, 287–292. Marques, J.M., Yzerbyt, V.Y, & Leyens, J.-Ph. (1988). Extremity of judgments towards ingroup members as a function of ingroup identification. European Journal of Social Psychology, 18, 1–16. Marris, P. (1996). The politics of uncertainty: Attachment in private and public life. London: Routledge. Martin, C.L., & Parker, S. (1995). Folk theories about sex and race differences. Personality and Social Psychology Bulletin, 21, 45–57.
352
REFERENCES
Martin, R., & Hewstone, M. (2001). Conformity and independence. In M.A.Hogg & R.S.Tindale (Eds.), Blackwell handbook of social psychology: Group processes (pp. 209–234). Oxford: Blackwell. Maslow, A.H. (1973a). The role of dominance in the social and sexual behavior of infra-human primates: I. Observations at Vilas Park Zoo. In Lowry (Ed.), Dominance, self-esteem, self-actualization: Germinal papers of Maslow (pp. 158, Belmont, CA: Wadsworth. Maslow, A.H. (1973b). The role of dominance in the social and sexual behavior of infra-human primates: II. An experimental determination of the behavior syndrome of dominance. In Lowry (Ed.), Dominance, self-esteem, selfactualization: Germinal papers of Maslow (pp. 18–46). Belmont, CA: Wadsworth. Maurer, K.L., Park, B., & Rothbart, M. (1995). Subtyping versus subgrouping processes in stereotype representation. Journal of Personality and Social Psychology, 69, 812–824. Mayr, E. (1982). The growth of biological thought: Diversity, evolution, and inheritance. Cambridge, MA: Harvard University Press. McCallum, D.M., Harring, K., Gilmore, R., Drenan, S., Chase, J., Insko, C.A., & Thibaut, J. (1985). Competition between groups and individuals. Journal of Experimental Social Psychology, 21, 301–320. McCauley, C.R., & Segal, M.E. (1987). Social psychology of terrorist groups. In C. Hendrick (Ed.), Review of personality and social psychology (Vol. 9, pp. 231–256). Beverly Hills, CA: Sage. McConahay, J.B. (1986). Modern racism, ambivalence, and the Modern Racism Scale. In J.F.Dovidio & S.L.Gaertner (Eds.), Prejudice, discrimination, and racism (pp. 91– 125). Orlando, FL: Academic Press. McConnell, A.R. (2001). Lay theories: Consequences for social judgments of individuals. Journal of Experimental Social Psychology, 37, 215–227. McConnell, A.R., Sherman, S.J., & Hamilton, D.L. (1994a). Illusory correlation in the perception of groups: An extension of the distinctiveness-based account. Journal of Personality and Social Psychology, 67, 414–429. McConnell, A.R., Sherman, S.J., & Hamilton, D.L. (1994b). On-line and memory-based aspects of individual and group target judgements. Journal of Personality and Social Psychology, 67, 173–185. McConnell, A.R., Sherman, S.J., & Hamilton, D.L. (1997). Target entitativity: Implications for information processing about individual and group targets. Journal of Personality and Social Psychology, 72, 750–762. McDougall, W. (1920/1973). The group mind. New York: Arno Press. McGarty, C. (1999). Categorization in social psychology. London: Sage. McGarty, C. (2001). Social identity theory does not maintain that identification produces bias, and self-categorization theory does not maintain that salience is identification: Two comments on Mummendey, Klink, and Brown. British Journal of Social Psychology, 40, 173–176. McGarty, C. (2002). Stereotype formation as category formation. In C.McGarty, V.Y. Yzerbyt, & R.Spears (Eds.) Stereotypes as explanations: The formation of meaningful beliefs about social groups (pp. 16–37). Cambridge: Cambridge University Press. McGarty, C., & de la Haye, A.-M. (1997). Stereotype formation: Beyond illusory correlation. In R.Spears, P.J.Oakes, N.Ellemers, & S.A.Haslam (Eds.), The social psychology of stereotyping and group life. (pp. 144– 70). Oxford: Blackwell. McGarty, C., & Groves, C. (2000). Illusory correlation and set size: Information loss and differentiated meaning. Unpublished manuscript, The Australian National University. McGarty, C., Haslam, S.A., Hutchinson, K.J., & Grace, D.M. (1995). Determinants of perceived consistency: The relationship between group entitativity and the meaningfulness of categories. British Journal of Social Psychology, 34, 237–256. McGarty, C., Haslam, S.A., Turner, J.C., & Oakes, P.J. (1993). Illusory correlation as accentuation of actual intercategory difference: Evidence for the effect with minimal stimulus information. European Journal of Social Psychology, 23, 391–410. McGarty, C., & Penny, R.E.C. (1988). Categorization, accentuation and social judgement. British Journal of Social Psychology, 22, 147–157. McGarty, C., & Turner, J.C. (1992). The effects of categorization on social judgment. British Journal of Social Psychology.31, 253– 268.
REFERENCES
353
McGarty, C., Yzerbyt, V.Y., & Spears, R. (2002). Stereotypes as explanations: The formation of meaningful beliefs about social groups. Cambridge: Cambridge University Press. McGrath, J.E. (1984). Groups: Interaction and performance. Englewood Cliffs, NJ: Prentice Hall. McGuire, W.J., & McGuire, C.V. (1982). Significant others in self-space: Sex differences in developmental trends in the self. In J.Suls (Ed.), Psychological perspectives on the self (Vol. 1, pp. 71–96). Hillsdale, NJ: Erlbaum. McIntyre, A., Lyons, A., Clark, A., & Kashima, Y. (2004). The microgenesis of culture: Serial reproduction as an experimental simulation of cultural dynamics. In M. Schaller & C.Crandall (Eds.), The psychological foundations of culture (pp. 227–258). Mahwah, NJ: Erlbaum. McNeil, W.H. (1986). Polyethnicity and national unity in world history. Toronto: University of Toronto Press. McPherson, M. (1983). An ecology of affiliation. American Sociological Review, 48, 519– 532. Medin, D.L. (1988). Social categorization: Structures, processes and purposes. In T.K. Srull & R.S.Wyer (Eds.), Advances in social cognition: A dual process model of impression formation (vol. 1). Hillsdale, NJ: Erlbaum. Medin, D.L., Goldstone, R.L., & Gentner, D. (1993). Respects for similarity. Psychological Review, 100, 254–278. Medin, D.L., & Ortony, A. (1989). Psychological essentialism. In S.Vosnaidou & A. Ortony (Eds.), Similarity and analogical reasoning (pp. 179–195). Cambridge: Cambridge University Press. Mendoza-Denton, R. (1999). Lay contextualism in stereotyping: Situational qualifiers of stereotypes in intuitive theories of dispositions [Abstract]. Dissertation Abstracts International, 60, 0412. Menon, T., Morris, M.W., Chiu, C.Y., Hong, Y.Y. (1999). Culture and the construal of agency: Attribution to individual versus group dispositions. Journal of Personality and Social Psychology, 76, 701–717. Messick, D.M., & Mackie, D.M. (1989). Intergroup relations. In M.R.Rosenzweig & L.W.Porter (Eds.), Annual review of psychology (Vol. 40, pp. 45–81). Palo Alto, CA: Annual Reviews. Middents, G.J. (1990). Psychological perspectives on enemy-making. Organization Development Journal, 8, 44–48. Mignerey, J.T., Rubin, R.B., & Gorden, W.I. (1995). Organizational entry: An investigation of newcomer communication behavior and uncertainty. Communication Research, 22, 54–85. Miller, D.T., & Prentice, D.A. (1996). The construction of social norms and standards. In E.T.Higgins & A.W.Kruglanski (Eds.), Social psychology: Handbook of basic principles (pp. 789–829). New York: Guilford Press. Miller, D.T., & Prentice, D.A. (1999). Some consequences of a belief in group essence: The category divide hypothesis. In D.A. Prentice & D.T.Miller (Eds.), Cultural divides: Understanding and overcoming group conflict (pp. 213–238). New York: Russell Sage Foundation. Miller, N., Gross, S., & Holtz, R. (1991). Social projection and attitudinal certainty. In J.Suls & T.A.Wills (Eds.), Social comparison: Contemporary theory and research (pp. 177–209). Hillsdale, NJ: Erlbaum. Miller, V.D., & Jablin, F.M. (1991). Information seeking during organizational entry: Influences, tactics, and a model of the process. Academy of Management Review, 16, 92–120. Mischel, W. (1997). Was the cognitive revolution just a detour on the road to behaviorism? On the need to reconcile situational control and personal control. In R.S.Wyer (Ed.), The automaticity of everyday life: Advances in social cognition (Vol. X, pp. 181–186). Mahwah, NJ: Erlbaum. Mischel, W., & Shoda, Y. (1995). A cognitive-affective system theory of personality: Reconceptualizing situations, dispositions, dynamics, and invariance in personality structure. Psychological Review, 102, 246– 268. Mohamed, A.A., Gardner, W.L., & Paolillo, J. G.P. (1999). A taxonomy of organizational impression management tactics. Advances in Competitiveness Research, 7, 108–130. Molden, D.C., Plaks, J.E., & Dweck, C.S. (2001). Lay theories and Person X Situation interactions: Processes of dispositional inference. Manuscript in preparation. Monteith, M.J., Sherman, J.W., & Devine, P. G. (1998). Suppression as a stereotype control strategy. Personality and Social Psychology Review, 2, 63–82. Moreland, R.L. (1987). The formation of small groups. In C.Hendrick (Ed.), Review of personality and social psychology (Vol. 8, pp. 80–110). Beverly Hills, CA: Sage. Moreland, R.L., Hogg, M.A., & Hains, S. (1994). Back to the future: Social psychological research on groups. Journal of Experimental Social Psychology, 30, 527–555.
354
REFERENCES
Moreland, R.L., & Levine, J.M. (1982). Socialization in small groups: Temporal changes in individual-group relations. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 15, pp. 137–192). New York: Academic Press. Moreland, R.L., & Levine, J.M. (1989). Newcomers and oldtimers in small groups. In P. B.Paulus (Ed.), Psychology of group influence (2nd ed., pp. 143–186). Hillsdale, NJ: Erlbaum. Moreland, R.L., & Levine, J.M. (1992a). Problem identification by groups. In S. Worchel, W.Wood, & J.A.Simpson (Eds.), Group process and productivity (pp. 17–47) Newbury Park, CA: Sage. Moreland, R.L., & Levine, J.M. (1992b). The composition of small groups. In E. Lawler, B.Markovsky, C.Ridgeway, & H.Walker (Eds.), Advances in group processes (Vol. 9, pp. 237–280). Greenwich, CT: JAI Press. Moreland, R.L., Levine, J.M., & Wingert, M. L. (1996). Creating the ideal group: Composition effects at work. In J.Davis & E. Witte (Eds.), Understanding group behavior (Vol. 2, pp. 11–35). Hillsdale, NJ: Erlbaum. Moreland, R.L., &: McMinn, J.G. (1999). Gone, but not forgotten: Loyalty and betrayal among ex-members of small groups. Personality and Social Psychology Bulletin, 25, 1476–1486. Morrison, E.W. (1993). Longitudinal study of the effects of information seeking on newcomer socialization. Journal of Applied Psychology, 78, 173–183. Moscovici, S, (1976). Social influence and social change. London: Academic Press. Moscovici, S. (1980). Toward a theory of conversion behavior. In L.Berkowitz (Ed.), Advances in experimental social psychology (Vol. 13, pp. 202–239). New York: Academic Press. Moscovici, S., & Zavalloni, M. (1969). The group as a polarizer of attitudes. Journal of Personality and Social Psychology, 12, 125–135, Moskowitz, G. (1993). Individual difference in social categorization-. The influence of personal need for structure on spontaneous trait inferences. Journal of Personality and Social Psychology, 65, 132–142. Mueller, C., & Dweck, C.S. (1998). Praise for intelligence can undermine children’s motivation and performance. Journal of Personality and Social Psychology, 75, 33–52. Mugny, G. (1982). The power of minorities. London: Academic Press. Mulder, M. (1977). The daily power game. Leiden: Martinus Nijhoff. Mullen, B. (1991). Group composition, salience, and cognitive representations: The phenomenology of being in a group. Journal of Experimental Social Psychology, 27, 297–323. Mullen, B., Brown, R., & Smith, C. (1992). Ingroup bias as a function of salience, rel evance, and status: An integration. European Journal of Social Psychology, 22, 103–122. Mullen, B., & Hu, L. (1989). Perceptions of ingroup and outgroup variability: A meta-analytic integration. Basic and Applied Social Psychology, 10, 233–252. Mullen, B., & Johnson, C. (1990). Distinctiveness-based illusory correlations and stereotyping: A meta-analytic integration. British Journal of Social Psychology, 29, 11–28. Mullin, B.-A., & Hogg, M.A. (1998). Dimensions of subjective uncertainty in social identification and minimal intergroup discrimination. British Journal of Social Psychology, 37, 345–365. Mullin, B.-A., & Hogg, M.A. (1999). Motivations for group membership: The role of subjective importance and uncertainty reduction. Basic and Applied Social Psychology, 21, 91–102. Murphy, G.L., & Medin, D.L. (1985). The role of theories in conceptual coherence. Psychological Review, 92, 289–316. Myers, D.G., & Bishop, G.D. (1970). Discussion effect on racial attitudes. Science, 169, 778–789. Naisbitt, J. (1984). Megatrends: Ten new directions transforming our lives. New York: Warner. Neuberg, S.L., & Fiske, S.T. (1987). Motivational influences on impression formation: Outcome dependency, accuracydriven attention, and individuating processes. Journal of Personality and Social Psychology, 53, 431–444. Neuberg, S.L., & Newson, J.T. (1993). Personal need for structure: Individual differences in the desire for simpler structure. Journal of Personality and Social Psychology, 65, 113–131. Ng, S.H. (1980). The social psychology of power. London: Academic Press. Ng, S.H. (1982). Power and intergroup discrimination. In H.Tajfel (Ed.), Social identity and intergroup relations (pp. 179–206). Cambridge: Cambridge University Press.
REFERENCES
355
Niemann, Y.R., Jennings, L., Rozelle, R.M., Baxter, J.C., & Sullivan, E. (1994). Use of free responses and cluster analysis to determine stereotypes of eight groups. Personality and Social Psychology Bulletin, 20, 379– 390. Nisbet, R. (1959). The decline and fall of social class. Pacific Sociological Review, 2, 11– 17. Nisbett, R.E., Krantz, D.H., Jepson, C., & Kunda, Z. (1983). The use of statistical heuristics in everyday intuitive reasoning. Psychological Review, 90, 339–363. Nisbett, R.E., & Kunda, Z. (1985). Perception of social distributions. Journal of Personality and Social Psychology, 48, 297–311. Nisbett, R.E., & Ross, L. (1980). Human inference: Strategies and shortcomings of social judgment. Englewood Cliffs, NJ. Prentice Hall. Nissen, M.J., Ross, J.L., Willingham, D.B., Mackenzie, T.B., Schacter, D.L. (1988). Memory and awareness in a patient with multiple personality disorder. Brain and Cognition, 8, 117–134. Nolan, M.A., Haslam, S.A., Spears, R., & Oakes, P.J. (1999). An examination of resource-based and fit-based theories of stereotyping under cognitive load and fit. European Journal of Social Psychology, 29, 641–663. Norman, W.T. (1963). Toward an adequate taxonomy of personality attributes: Replicated factor structures in peer nomination personality ratings. Journal of Abnormal and Social Psychology, 66, 574–583. Nosek, B., & Banaji, M.R. (2001). The go/no-go association task. Social Cognition, 19, 625–666. Nosofsky, R.M., Palmeri, T.J., & McKinley, S.C. (1994). Rule-plus-exception model of classification learning. Psychological Review, 101, 53–79. Oakes, P.J. (1987). The salience of social categories. In J.C.Turner, M.A.Hogg, P.J. Oakes, S.D.Reicher, & M.S.Wetherell (Eds.), Rediscovering the social group: A self-categorization theory (pp. 117–141). Oxford: Blackwell. Oakes, P.J. (1996). The categorization process: Cognition and the group in the social psychology of stereotyping. In W.P.Robinson (Ed.), Social groups and identities: Developing the legacy of Henri Tajfel (pp. 95–120). Oxford: Butterworth-Heinemann. Oakes, P.J. (2000). The root of all evil in intergroup relations? Unearthing the categorization process. In R.Brown & S.Gaertner (Eds.) Blackwell handbook in social psychology: Vol. 4: Intergroup processes. Oxford: Blackwell. Oakes, P.J., Haslam, S.A., Morrison, B., & Grace, D. (1995). Becoming an in-group: Reexamining the impact of familiarity on perceptions of group homogeneity. Social Psychology Quarterly, 58, 52–61. Oakes, P.J., Haslam. S.A., & Reynolds, K.J. (1999) Social categorization and social context: Is stereotype change a matter of information or of meaning? In D.Abrams & M. A.Hogg (Eds.), Social identity and social cognition (pp. 55–79). Oxford: Blackwell. Oakes, P.J., Haslam, S.A., & Turner, J.C. (1994) Stereotyping and social reality. Oxford: Blackwell. Oakes, P.J., Reynolds, K.J., Haslam, S.A., & Turner, J.C. (1999). Part of life’s rich tapes-try: Stereotyping and the politics of intergroup relations. In E.J.Lawler, S.Thye, M. W.Macy & H.A.Walker (Eds.), Advances in group processes (Vol. 16,125–160). Stamford, CT: JAI Press. Oakes, P.J., & Turner, J.C. (1980). Social categorization and intergroup behaviour: Does minimal intergroup discrimination make social identity more positive? European Journal of Social Psychology, 10, 295–301. Oakes, P.J. & Turner, J.C. (1990). Is limited information processing capacity the cause of social stereotyping? In W.Stoebe & M. Hewstone (Eds.), European review of social psychology (Vol. 1, pp. 111–135). Chichester, UK: Wiley. Oakes, P.J., Turner, J.C., & Haslam, S.A. (1991). Perceiving people as group members: The role of fit in the salience of social categorizations. British Journal of Social Psychology, 30, 125–144. O’Laughlin, M.J., & Malle, B. (2002). How people explain actions performed by groups and individuals. Journal of Personality and Social Psychology, 82, 33–48. Oots, K.L. (1989). Organizational perspectives on the formation and dissolution of terrorist groups. Terrorism, 12, 139–152. Operario, D., & Fiske, S.T. (1998) Racism equals power plus prejudice. In J.L. Eberhardt and S.T.Fiske (Eds), Confronting racism: The problem and the response (pp. 33–53). Thousand Oaks, CA: Sage.
356
REFERENCES
Operario, D., & Fiske, S.T. (1999). Integrating social identity and social cognition: A framework for bridging diverse perspectives. In D.Abrams & M.A.Hogg (Eds.), Social identity and social cognition (pp. 26–54). Maiden, MA: Blackwell. Ostrom, T.M. (1984). The sovereignty of social cognition. In R.S.Wyer, Jr., & T.K. Srull. (Eds.), Handbook of social cognition (Vol. 1, pp. 1–38). Hillsdale, NJ: Erlbaum. Ostrom, T.M., Carpenter, S.L., Sedikides, C., & Li, F. (1993). Differential processing of in-group and out-group information. Journal of Personality and Social Psychology, 64, 21–34. Ostrom, T.M., & Sedikides, C. (1992). Outgroup homogeneity effects in natural and minimal groups. Psychological Bulletin, 112, 536–552. Otten, S., & Wentura, D. (2001). Self-anchoring and in-group favoritism: An individual profile analysis. Journal of Experimental Social Psychology, 37, 525–532. Overbeck, J.R., & Park, B. (2001). When power does not corrupt: Superior individuation processes among powerful perceivers. Journal of Personality and Social Psychology, 81, 549–565. Park, B., & Hastie, R. (1987). Perception of variability in category development: Instance- versus abstraction-based stereotypes, Journal of Personality and Social Psychology, 53, 621–635. Park, B., & Judd, C.M. (1990). Measures and models of perceived group variability. Journal of Personality and Social Psychology, 59, 173–191. Park, B., Judd, C.M., & Ryan, C.S. (1991). Social categorization and the representation of variability information. In W. Stroebe & M.Hewstone (Eds.), European review of social psychology (Vol. 2, pp. 211–245). Chichester: Wiley. Park, B., & Rothbart, M. (1982). Perception of out-group homogeneity and levels of social categorization: Memory for the subordinate attributes of in-group and out-group members. Journal of Personality and Social Psychology, 42, 1051–1068. Park, B., Ryan, C.S., & Judd, C.M. (1992). Role of meaningful subgroups in explaining differences in perceived variability for in- groups and out-groups. Journal of Personality and Social Psychology, 63, 553–567. Park, B., Wolsko, C., & Judd, C.M. (2001). Measurement of subtyping in stereotype change. Journal of Experimental Social Psychology, 37, 325–332. Parks, M.R., & Eggert, L.L. (1991). The role of social context in the dynamics of personal relationships. In W.H.Jones & D.Perlman (Eds.), Advances in personal relationships (Vol. 2, pp 1–34). London: Jessica Kingsley. Parks, M.R., Stan, C.M., & Eggert, L.L. (1983). Romantic involvement and social network involvement. Social Psychology Quarterly, 46, 116–131. Patterson, M.L. (1985). Social influence and nonverbal exchange. In S.Ellyson & J.F. Dovidio (Eds.), Power, dominance, and nonverbal behavior (pp. 207–217), New York: Springer-Verlag, Paulhus, D.L. (1993). Bypassing the will: The automatization of affirmations. In D.M. Wegner & J.W.Pennebaker (Eds.), Handbook of mental control (pp. 573–587). Englewood Cliffs, NJ: Prentice Hall Pavelchak, M.A., More land, R.L., & Levine, J.M. (1986), Effects of prior group memberships on subsequent reconnaissance activities. Journal of Personality and Social Psychology, 50, 56–66. Payne, J.W., Bettman, J.R., & Johnson, E.J. (1993). The adaptive decision maker. Cambridge: Cambridge University Press. Pemberton, M.B., Insko, C.A., & Sehopler, J. (1996). Memory for and experience of differential competitive behavior of individuals and groups. Journal of Personality and Social Psychology, 71, 953–966. Pepper, S. (1942), World hypotheses. Berkeley: University of California Press. Perdue, C.W., Dovidio, J.F., Gurtman, M.B., & Tyler, R.B. (1990). Us and them: Social categorization and the process of intergroup bias. Journal of Personality and Social Psychology, 59, 475–486. Pettigrew, T.F. (1979). The ultimate attribution error: Extending Allport’s cognitive analysis of prejudice. Personality and Social Psychology Bulletin, 5, 461–476. Pettigrew, T.F. (1998). Intergroup contact theory. Annual Review of Psychology, 49, 65–85. Pettigrew, T., & Martin, J. (1987). Shaping the organizational context for Black American inclusion. Journal of Social Issues, 43, 41–78.
REFERENCES
357
Petty, R.E., & Wegener, D.T. (1998). Attitude change: Multiple roles for persuasion. In D.T.Gilbert, S.T.Fiske & G.Lindzey (Eds.), The handbook of social psychology (4th ed., Vol.1, pp. 323–390). New York: McGraw-Hill. Pfeffer, J. (1981). Management as symbolicaction: The creation and maintenance of organizational paradigms. In L.L.Cummings & B.M.Staw (Eds.), Research in organizational behavior (Vol. 3, pp. 1–52). Greenwich, CT: JAI Press. Phillips, J.M. (1998). Effects of realistic job previews on multiple organizational outcomes: A meta-analysis. Academy of Management Journal, 41, 673–690. Pichevin, M.-F., & Hurtig, M.-C. (1996). Describing men, describing women: Sex membership salience and numerical distinctiveness. European Journal of Social Psychology, 26, 513–522. Pickett, C.L. (2001). The effects of entitativity beliefs on implicit comparisons between group members. Personality and Social Psychology Bulletin, 27, 515–525. Pickett, C.L., & Brewer, M.B. (2001). Assimilation and differentiation needs as motivational determinants of perceived in-group and out-group homogeneity. Journal of Experimental Social Psychology, 37, 341–348. Pinter, B., & Insko, C.A, (2003). Implicit distrust in the absence of context: Generality of the negative outgroup shema. Submitted for publication. Pittman, T.S. (1993). Control motivation and attitude change. In G.Weary, F.Gleicher, & K.L.Marsh (Eds.), Control motivation and social cognition (pp. 157–175). New York: Springer-Verlag. Pittman, T.S., & Pittman, N.L. (1980). Deprivation of control and the attribution process. Journal of Personality and Social Psychology, 25, 465–480. Plaks, J.E., Grant-Pillow, H.A., & Dweck, C. S. (2002). When is inconsistent person information a threat? Violation of lay theories of personality and the motivation for prediction and control. Manuscript under review view. Plaks, J.E., Shafer, J.L., & Shoda, Y. (2003). Perceiving individuals and groups as coherent: How do perceivers make sense of variable information? Social Cognition, 21, 26– 60. Plaks, J.E., Stroessner, S.J., Dweck, C.S., & Sherman, J. (2001). Person theories and attention allocation: Preferences for stereotypic vs. counterstereotypic information. Journal of Personality and Social Personality, 80, 876–893. Plant, E.A., & Devine, P.G. (1998). Internal and external motivation to respond without prejudice. Journal of Personality and Social Psychology, 75, 811–832. Plato (1891). Republic. In B.Jowett (Ed., & Trans.), The dialogues of Plato (Vol.3, pp. 1– 338). Oxford: Oxford University Press. Ponse, B. (1976). Secrecy in the lesbian world. Urban Life, 5, 313–337. Popper, K. (1982). Unended quest: An intellectual autobiography. LaSalle, IL: Open Court Press. Postmes, T., & Spears, R. (1999). Anonymity in computer-mediated communication: Different groups have different SIDE effects. Manuscript submitted for publication. Postmes, T., Spears, R., Sakhel, K., & de Groot, D. (2001). Social influence in computer-mediated groups: The effects of anonymity on social behavior. Personality and Social Psychology Bulletin, 27, 1243–1254 Powell, G.N., & Foley, S. (1999). Romantic relationships in organizational settings. In G. N.Powell (Ed.), Handbook of gender and work (pp. 281–304). Thousand Oaks, CA: Sage. Powlishta, K.K. (1995). Intergroup processes in childhood: Social categorization and sex role development. Developmental Psychology, 31, 781–788. Prentice, D.A., Miller, D.T., & Lightdale, J. R. (1994). Asymmetries in attachments to groups and to their members: Distinguishing between common identity and commonbond groups. Personality and Social Psychology Bulletin, 20, 484–493. Pryor, J.B. (1986). The influence of different encoding sets upon the formation of illusory correlations and group impressions. Personality and Social Psychology Bulletin, 12, 216–226. Putnam, H. (1975). Mind, language, and reality. Cambridge: Cambridge University Press. Quattrone, G.A., & Jones, E.E. (1980). The perception of variability within ingroups and outgroups: Implications for the law of small numbers. Journal of Personality and Social Psychology, 38, 141–152.
358
REFERENCES
Queller, S., Mackie, D. M., & Stroessner, S.J. (1996). Ameliorating some negative effects of positive mood: Encouraging happy people to perceive intragroup variability. Journal of Experimental Social Psychology, 32, 361– 386. Quinn, R.E. (1977). Coping with Cupid: The formation, impact, and management of romantic relationships in organizations. Administrative Science Quarterly, 22, 30–45. Rabbie, J.M., & Horwitz, M. (1969). Arousal of ingroup-outgroup bias by a chance win or loss. Journal of Personality and Social Psychology, 13, 269–277. Rabbie, J.M., Schot, J.C., & Visser, L. (1989) Social identity theory: A conceptual and empirical critique from the perspective of a Behavioural Interaction Model. European Journal of Social Psychology, 19, 171–202. Ramsey, S.L., Lord, C.G., Wallace, D.S., & Pugh, M.A. (1994). The role of subtypes in attitudes towards superordinate social categories. British Journal of Social Psychology, 33, 387–403. Rapley, M. (1998). ‘Just an ordinary Australian’: Self-categorization and the discursive construction of facticity in ‘new racist’ political rhetoric. British Journal of Social Psychology, 37, 325–344. Rapoport, A., & Bornstein, G. (1987). Intergroup competition for the provision of binary public goods. Psychological Review, 94, 291–299. Reeder, G.D., & Brewer, M.B. (1979). A schematic model of dispositional attribution in interpersonal perception. Psychological Review, 86, 61–79. Reicher, S.D. (1996). Social identity and social change: Rethinking the context of social psychology. In P.Robinson (Ed.) Social groups and identities: Developing the legacy of Henri Tajfel (pp. 317–336). Oxford: ButterworthHeinemann. Reicher, S.D. (2001) Studying psychology studying racism. In M.Augoustinos & K.J. Reynolds (Eds.), Understanding prejudice, racism, and social conflict. London: Sage. Reicher, S.D. (in press). The context of social identity: Domination, resistance, and change. Political Psychology. Reicher, S.D., & Hopkins, N. (2001). Self and Nation. London: Sage. Reicher, S.D., Hopkins, N., & Condor, S. (1997), Stereotype construction as a strategy of social influence. In R.Spears, P.J. Oakes, N.Ellemers, & S.A.Haslam (Eds.), The social psychology of stereotyping and group life (pp. 94–118). Oxford: Blackwell. Reicher, S.D., Spears, R., & Postmes, T. (1995). A social identity model of deindi-viduation phenomena. In W. Stroebe & M. Hewstone (Eds.), European review of social psychology (Vol. 6, pp. 161–198). Chichester, UK: Wiley. Reid, S.A., & Hogg, M.A. (2001). Uncertainty reduction, self-enhancement, and social identification. Manuscript submitted for publication, University of California, Santa Barbara. Reis, H.R., & Wheeler, L. (1991). Studying social interaction with the Rochester Interaction Record. In M.P.Zanna (Ed.), Advances in experimental social psychology (Vol. 24, pp. 269–318). New York: Academic Press. Reynolds, K.J., & Oakes, P.J. (2000). Variability in impression formation: Investigating the role of motivation, capacity, and the categorization process. Personality and Social Psychology Bulletin, 26, 355–373. Reynolds, K.J., Oakes, P.J., & Haslam, S.A. (1999). To think, feel, and act as a collective: The role of intergroup versus intragroup self-other comparisons. Invited paper presented at the Homogeneity and Entitativity EAESP Small Group Meeting in Louvain-la-Neuve, Belgium, 4–6 July, Reynolds, K.J., Oakes, P.J., Haslam, S.A., Nolan, M.A., & Dolnik, L. (2000) Responses to powerlessness: Stereotypes as an instrument of social conflict. Group Dynamics: Theory, Research and Practice, 4, 275– 290. Reynolds, K.J., Oakes, P.J., Haslam. S.A., Spears, R., & Wegener, R. (1998). Does having power always lead to stereotyping? Investigating cognitive capacity and group-based accounts of the stereotyping process [Abstract only]. Australian Journal of Psychology, 50, 32. Reynolds, K.J., & Platow, M. (2003). On the social psychology of power and powerlessness: Social power as a symptom of organizational division. In S.A.Haslam, D.van Knippenberg, N.Ellemers, & M.Platow (Eds.) Social identity at work: Developing theory for organizational practice (pp. 173– 188). Philadelphia: Psychology Press. Rhee, E., Uleman, J., Lee, H.K., & Roman, R.J. (1995). Spontaneous self-descriptions and ethnic identities in individualistic and collectivistic cultures. Journal of Personality and Social Psychology, 69, 142–152.
REFERENCES
359
Rhodewalt, F. (1994). Conceptions of ability, achievement goals, and individual differences in self-handicapping behavior: On the application of lay theories. Journal of Personality, 62, 67–85. Richards, Z., & Hewstone, M. (2001). Subtyping and subgrouping: Processes for the prevention and promotion of stereotype change. Personality and Social Psychology Review, 5, 52–73. Richardson, L. (1986, February). Another world. Psychology Today, 20, 22–27. Ridgeway, C.L., & Diekema, D. (1992). Are gender differences status differences? In C.L.Ridgeway (Ed.), Gender, interaction, and inequality (pp. 157–180). New York: Springer-Verlag. Ridley, M. (1996). The origins of virtue. London: Penguin. Robinson, S.L., & Rousseau, D.M. (1994). Violating the psychological contract: Not the exception but the norm. Journal of Organizational Behavior, 15, 245–259. Rodgers, J., Hamilton, D.L., & Sherman, S.J. (2003). Cognitive representations and beliefs about task groups and social categories: Implications for entitativity and stereotyping. Unpublished manuscript, University of California, Santa Barbara. Rogier, A., & Yzerbyt, V.Y. (1999). Social attribution, correspondence bias, and the emergence of stereotypes. Swiss Journal of Psychology, 54, 233–240. Rokeach, M. (1948). Generalized mental rigidity as a factor in ethnocentrism. Journal of Abnormal and Social Psychology, 43, 259– 278. Rokeach, M. (1960). The open and closed mind. New York: Basic Books. Rommetveit, R. (1974). On message structure: A framework for the study of language and communication. New York: Wiley. Rosch, E. (1975). Cognitive representations of semantic categories. Journal of Experimental Psychology: General, 104, 192–233. Rosch, E., Mervis, C.B., Gray, W.D., Johnson, D.M., & Boyes-Braem, P. (1976). Basic objects in natural categories. Cognitive Psychology, 8, 382–439. Rosenberg, S., & Sedlak, A. (1972). Structural representations of implicit personality theory. In L.Berkowitz (Ed.), Advances in experimental social psychology (Vol. 6, pp. 235–297). New York: Academic Press. Rosenblatt, A., Greenberg, J., Solomon, S., Pyszczynski, T., & Lyon, D. (1989). Evidence for terror management theory I: The effects of mortality salience on reactions to those who violate or uphold cultural values. Journal of Personality and Social Psychology, 57, 279–301. Ross, J., & Staw, B.M. (1993). Organizational escalation and exit: Lessons from the Shoreham Nuclear Power Plant. Academy of Management Journal, 36, 701–732. Ross, L.D., Amabile, T.M., & Steinmetz, J. L. (1977). Social roles, social control, and biases in social perception processes. Journal of Personality and Social Psychology, 35, 485–494. Ross, L.D., & Nisbett, R.E. (1991). The person and the situation: Perspectives of social psychology. New York: McGraw-Hill. Rothbart, M. (1981). Memory processes and social beliefs. In D.L.Hamilton (Ed.), Cogntive processes in stereotyping and intergroup behavior (pp. 145–181). Hillsdale, NJ: Erlbaum. Rothbart, M., & John, O.P. (1985). Social categorization and behavioral episodes: A cognitive analysis of the effects of intergroup contact. Journal of Social Issues, 41, 81–104. Rothbart, M., & Lewis, S. (1988). Inferring category attributes from exemplar attributes: Geometric shapes and social categories. Journal of Personality and Social Psychology, 55, 861–872. Rothbart, M., & Park, B. (1986). On the confirmability and disconfirmability of trait concepts. Journal of Personality and Social Psychology, 50, 131–142. Rothbart, M., Peters, E., & Buckingham, S. (2002). Do lesbians drink orange juice? Category strength and mundane behaviors. Unpublished manuscript, University of Oregon, Eugene. Rothbart, M., Sriram, N., & Davis-Stitt, C. (1996). The retrieval of typical and atypical category members. Journal of Experimental Social Psychology, 32, 1–29. Rothbart, M., & Taylor, M. (1992). Category labels and social reality: Do we view social categories as natural kinds? In G.R. Semin & K.Fiedler (Eds.), Language, interaction and social cognition (pp. 11–36). London: Sage.
360
REFERENCES
Rothgerber, H. (1997). External intergroup threat as an antecedent to perceptions of in-group and out-group homogeneity. Journal of Personality and Social Psychology, 73, 1206–1211. Rubin, M., & Hewstone, M. (1998). Social identity theory’s self-esteem hypothesis: A review and some suggestions for clarification. Personality and Social Psychology Review, 2, 40–62. Rubin, M., Hewstone, M., & Voci, A. (2001). Stretching the boundaries: Strategic perceptions of intragroup variability. European Journal of Social Psychology, 31, 413–429. Rubin, M., Richards, Z., & Hewstone, M. (2000). Unpublished raw data, Cardiff University. Ruscher, J.B., & Fiske, S.T. (1990). Interpersonal competition can cause individuating processes. Journal of Personality and Social Psychology, 58, 832–843. Ryan, C.S., & Bogart, L.M. (1997). Development of new group members’ in-group and out-group stereotypes: Changes in perceived group variability and ethnocentrism. Journal of Personality and Social Psychology, 73, 719–32. Ryan, C.S., & Bogart, L.M. (2001). Longitudinal changes in the accuracy of new group members’ in-group and outgroup stereotypes. Journal of Experimental Social Psychology, 37, 118–133. Ryan, C.S., Bogart, L.M., & Vender, J.P. (2000). Effects of perceived group variability on the gathering of information about individual group members. Journal of Experimental Social Psychology, 36, 90–101. Ryan, C.S., Judd, C.M., & Park, B. (1996). Effects of racial stereotypes on judgments of individuals: The moderating role of perceived group variability. Journal of Experimental Social Psychology, 32, 71–103. Ryan, C.S., Park, B., & Judd, C.M. (1996). Assessing stereotype accuracy: Implications for understanding the stereotyping process. In C.N.Macrae, C.Stangor, & M.Hewstone (Eds.), Stereotypes and stereotyping (pp. 121–157). New York: Guilford Press. Ryan, C.S., Robinson, D.R., & Hausmann, L. R.M. (2001). Stereotyping among providers and consumers of public mental health services: The role of perceived group variability. Behavior Modification, 25, 407–443. Ryder, A.G., Kafka, J.S., & Olsen, D.H. (1971). Separating and joining influences in courtship and early marriage. Journal of Orthopsychiatry, 41, 450–464. Sacchi, E., & Castano, E. (2002). Entitative is beautiful: The importance of perceiving the ingroup as a real entity. European Association of Experimental Social Psychology (EAESP), 13th general meeting, San Sebastian, Spain, June 2002. Sachdev, I., & Bouhris, R.Y. (1984). Minimal majorities and minorities. European Journal of Social Psychology, 14, 35–52. Sachdev, I., & Bourhis, R.Y. (1991). Power and status differentials in minority and majority group relations. European Journal of Social Psychology, 21, 1–24. Saks, A.M. (1995). Longitudinal field investigation of the moderating and mediating effects of self-efficacy on the relationship between training and newcomer adjustment. Journal of Applied Psychology, 80, 211–25. Saks, A.M., & Ashforth, B.E. (1997). Organizational socialization: Making sense of the past and present as a prologue for the future. Journal of Vocational Behavior, 51, 234–279. Saltz, E., & Medow, M.L. (1971). Concept conservation in children: The dependence of belief systems on semantic representation. Child Development, 42, 1533–1542. Sandelands, L., & St. Clair, L. (1993). Toward an empirical concept of group. Journal for the Theory of Social Behavior, 23, 423–458. Sani, F., & Reicher, S. (2000). Contested identities and schisms in groups: Opposing the ordination of women as priests in the Church of England. British Journal of Social Psychology, 39, 95–112. Scadding, J.G. (1996). Essentialism and nominalism in medicine: Logic of diagnosis in disease terminology. The Lancet, 348, 594– 596. Scarberry, N.C., Ratcliff, C.D., Lord, C.G., Lanicek, D.L., & Desforges, D.M. (1997). Effects of individuating information on the generalization part of Allport’s contact hypothesis. Personality and Social Psychology Bulletin, 23, 1291–1299. Schachter, S. (1951). Deviation, rejection and communication. Journal of Abnormal and Social Psychology, 46, 190–207.
REFERENCES
361
Scheepers, D., Spears, R., Doosje, B., & Manstead, A.S.R. (2003) Integrating identity and instrumental approaches to intergroup differentiation: Different contexts, different motives. Personality and Social Psychology Bulletin, 28, 1455–1467. Schlenker, B.R., & Britt, T.W. (1999). Beneficial impression management: Strategically controlling information to help friends. Journal of Personality and Social Psychology, 76, 559–573. Schlenker, B.R., Dlugolecki, D.W., & Doherty, K. (1994). The impact of self-presentations on self-appraisals and behavior: The roles of commitment and biased scanning. Personality and Social Psychology Bulletin, 20, 20–33. Schlenker, B.R., & Weigold, M.F. (1992). Interpersonal processes involving impression management and regulation. Annual Review of Psychology, 43, 133–168. Schmitt, M.T., & Branscombe, N.R. (2001). The meaning and consequences of perceived discrimination in disadvantaged and privileged social groups. In W.Stroebe & M. Hewstone (Eds.), European review of social psychology (Vol. 12, pp. 167–129). Chichester, UK: Wiley. Schneider, A., & Ingram, H. (1993). Social construction of target populations: Implications for politics and policy. American Political Science Review, 87, 334–347. Schneider, D.J. (1973) Implicit personality theory: A review. Psychological Bulletin, 79, 294–309. Schopler, J., Insko, C.A., Currey, D., Smith, S., Brazil, D., Riggins, T, Gaertner, L., et al. (1994). The survival of a cooperative tradition in the intergroup discontinuity context. Motivation and Emotion, 18, 301–315. Schopler, J., Insko, C.A., Drigotas, S.M., Wieselquist, J., Pemberton, M., & Cox C. (1995). The role of identifiability in the reduction of interindividual-intergroup discontinuity. Journal of Experimental Social Psychology, 31, 553–574. Schopler, J., Insko, C.A., Graetz, K.A., Drigotas, S.M., Smith, V.A. (1991). The generality of the individual-group discontinuity effect: Variations in positivity-negativity of outcomes, players’ relative power, and the magnitude of outcomes. Personality and Social Psychology Bulletin, 17, 612–624. Schopler, J., Insko, C.A., Graetz, K.A., Drigotas, S.M., Smith, V.A., & Dahl, K.A. (1993). Individual-group discontinuity: Further evidence for mediation by fear and greed. Personality and Social Psychology Bulletin, 19, 419–431. Schopler, J., Insko, C.A., Wieselquist, J., Pemberton, M.B., Witcher, B., Kozar, R., et al. (2001). When groups are more competitive than individuals: The domain of the discontinuity effect. Journal of Personality and Social Psychology, 80, 632–644. Schubert, T.W, & Otten, S. (in press). Overlap of self, ingroup, and outgroup: Pictorial measures of self-categorization. Self and Identity. Schuman, H., Steeh, C., Bobo, L., & Krysan, M. (1997). Racial attitudes in America: Trends and interpretations. Cambridge, MA: Harvard University Press. Scott, W.A. (1969). The structure of natural cognitions. Journal of Personality and Social Psycholoty, 12, 261–278. Searle, J.R. (1995). The construction of social reality. London: Penguin Books. Searle, J.R. (1998). Mind, language, and society. New York: Basic Books. Sedikides, C. (1997). Differential processing of ingroup and outgroup information: The role of relative group status in permeable boundary groups. European Journal of Social Psychology, 27, 121–144. Sedikides, C., & Brewer, M.B. (2001). Indi vidual self, relationship, collective self. Philadelphia: Psychology PRess. Sedikides, C., & Ostrom, T.M. (1993). Perceptions of group variability: Moving from an uncertain crawl to a purposeful stride. Social Cognition, 11, 165–174. Sedikides, C., & Skowronski, J.J. (1997). The symbolic self in evolutionary context. Personality & Social Psychology Review, 1, 80– 102. Sedikides, C., & Strube, M.J. (1995). The multiply motivated self. Personality and Social Psychology Bulletin, 21, 1330–1335. Sedikides, C., & Strube, M.J. (1997). Self-evaluation: To thine own self be good, to thine own self be sure, to thine own self be true, and to thine own self be better. In M. P.Zanna (Ed.), Advances in experimental social psychology (vol 29, pp. 209–296). New York: Academic Press. Seltzer, J.A. (1994). Consequences of marital dissolution for children. Annual Review of Sociology, 20, 235–266.
362
REFERENCES
Semin, G., & Fiedler, K. (1991). The Linguistic Category Model, its bases, applications and range. In W.Stroebe & M.Hewstone (Eds.), European review of social psychology (Vol. 2, pp. 1–30). Chichester, UK: Wiley. Sen, A.K. (1985). Commodities and capabilities: Lectures in economics policy (Vol. 7). New York: Elsevier Science. Shaw, D.S. (1992). The effects of divorce on children’s adjustment. In M.E. Procidiano & C.B.Fisher (Eds.), Contemporary families: A handbook for school professionals (pp. 179–205). New York: Teachers College Press. Shaw, M.E. (1976). Group dynamics: The psychology of small group behavior. New York: McGraw-Hill. Sherif, M. (1936). The psychology of social norms. New York: Harper & Row. Sherif, M. (1967). Group conflict and co-operation: Their social psychology. London: Routledge and Kegan Paul. Sherif, M., Harvey, O.J., White, B.J., Hood, W.R., & Sherif, C.W. (1961). Intergroup conflict and cooperation: The Robbers Cave experiments. Norman, OK: Institute for Group Relations. Sherif, M., Harvey, O.J., White, B.J., Hood, W.R., & Sherif, C.W. (1988). The Robbers Cave experiment: Intergroup conflict and cooperation. Middletown, CT: Wesleyan University Press. Sherif, M., & Hovland, C.I. (1961). Social judgment: Assimilation and contrast effects in communication and attitude change. New Haven, CT: Yale University Press. Sherman, J.W. & Frost, L.A. (2000). On the encoding of stereotype-relevant information under cognitive load. Personality and Social Psychology Bulletin, 26, 26–34. Sherman, J.W., Klein, S.B., Laskey, A., & Wyer, N.A. (1998). Intergroup bias in group judgment processes: The role of behavioral memories. Journal of Experimental Social Psychology 34, 51–65. Sherman, J.W., Lee, A.Y., Bessenoff, G.R., & Frost, L.A. (1998). Stereotype efficiency reconsidered: Encoding flexibility under cognitive load. Journal of Personality and Social Psychology, 75, 589–606. Sherman, S.J., Castelli, L., & Hamilton, D.L. (2002). The spontaneous use of a group typology as an organizing principle in memory. Journal of Personality and Social Psychology, 82, 328–342. Sherman, S.J., Hamilton, D.L., & Lewis, A. C. (1999). Perceived entitativity and the social identity value of group memberships. In D.Abrams & M.A.Hogg (Eds.), Social identity and social cognition (pp. 80–110). Maiden, MA: Blackwell. Showers, C. (1992). Compartmentalization of positive and negative self-knowledge: Keeping bad apples out of the bunch. Journal of Personality and Social Psychology, 62, 1036– 1049. Shure, G.H., Meeker, R.J., & Hansford, E.A. (1965). The effectiveness of pacifist strategies in bargaining games. Journal of Conflict Resolution, 9, 106–117. Sidanius, J., & Pratto, F. (1999). Social dominance: An intergroup theory of social hierarchy and oppression. New York: Cambridge University Press. Sillars, A., Shellen, W., McIntosh, A., & Pomegranate, M. (1997). Relational characteristics of language: Elaboration and differentiation in marital conversations. Western Journal of Communication, 61, 403–422. Silver, M., Conte, R., Miceli, M., & Poggi, I. (1986). Humiliation: Feeling, social control, and the construction of identity. Journal for the Theory of Social Behavior, 16, 269–283. Simon, B. (1992a). The perception of in-group and out-group homogeneity: Reintroducing the intergroup context. In W.Stroebe & M. Hewstone et al. (Eds.), European review of social psychology (Vol. 3, pp. 1–30). Chichester, UK: Wiley. Simon, B. (1992b). Intragroup differentiation in terms of in-group and out-group attributes. European Journal of Social Psychology, 22, 407–413. Simon, B. (1993). On the asymmetry in the cognitive construal of in-group and out-group: A model of egocentric social categorization. European Journal of Social Psychology, 23, 131–147. Simon, B. (1998a). Individuals, groups, and social change: On the relationship between individual and collective selfinterpretations and collective action. In C.Sedikides, J. Schopler, & C.Insko (Eds.), Intergroup cognition and intergroup behavior. Mahwah, NJ: Erlbaum. Simon, B. (1998b). The self in minority-majority contexts. In W.Stroebe & M. Hewstone (Eds.) European review of social psychology (Vol. 9, pp. 1–30). Chichester, UK: Wiley. Simon, B. & Brown, R. (1987). Perceived intragroup homogeneity in minority-majority contexts. Journal of Personality and Social Psychology, 53, 703–711.
REFERENCES
363
Simon, B., Glassner-Bayerl, B., & Straten-werth, I. (1991). Stereotyping and self-stereotyping in a natural intergroup context: The case of heterosexual and homosexual men. Social Psychology Quarterly, 54, 252– 266. Simon, B., & Hamilton, D.L. (1994). Self-stereotyping and social context: The effects of relative in-group size and ingroup status. Journal of Personality and Social Psychology, 66, 699–711. Simon, B., & Klandermans, B. (2001). Politicized collective identity: A social psychological analysis. American Psychologist, 56, 319–331. Simon, B., & Mummendey, A. (1990). Perceptions of relative group size and group homogeneity: We are the majority and they are all the same. European Journal of Social Psychology, 20, 351–356. Simon, B., & Pettigrew, T.F. (1990). Social identity and perceived group homogeneity: Evidence for the in-group homogeneity effect. European Journal of Social Psychology, 20, 269–286. Sindic, D., Castano, E., & Reicher, S.D. (2001). Les dynamiques identitaires et le processus d’intégration européenne. Etudes Internationales, 32, 425–454. Skinner, E.A. (1995). Perceived control, motivation, and coping. London: Sage. Smith, A.D. (1995). Nations and nationalism in a global era. Oxford: Blackwell Smith, E.E., & Medin, D.L. (1981). Categories and concepts. Cambridge, MA: Harvard University Press. Smith, E.R. (1991) Illusory correlation in a simulated exemplar-based memory. Journal of Experimental Social Psychology, 27, 107–123. Smith, E.R. (1998). Mental representations and memory. In D.T.Gilbert, S.T.Fiske, & G.Lindzey (Eds.), The handbook of social psychology (4th ed., pp. 391–445). Boston: McGraw-Hill. Smith, E.R. (in press). Overlapping metal representations of self and group: Evidence and implications. In J.P.Forgas and K.D.Williams (Eds.), The social self: Cognitive, interpersonal, and intergroup perspectives. Philadelphia: Psychology Press. Smith, E.R., Coats, S., & Murphy, J. (2001). The self and attachment to relationship partners and groups: Theoretical parallels and new insights. In C.Sedikides & M.B. Brewer (Eds). Individual self, relational self, collective self (pp. 109–124) Philadelphia: Psychology Press. Smith, E.R., & Zárate, M.A. (1992). Exemplar-based model of social judgment. Psychological Review, 99, 3–21. Snyder, C.R., & Fromkin, H.L. (1980). Uniqueness: The human pursuit of difference. New York: Plenum Press. Solomon, S., Greenberg, J., & Pyszcsynzski, T. (1991). A terror management theory of social behavior: The psychological functions of self-esteem and cultural worldviews. In P.M.Zanna (Ed.), Advances in experimental social psychology (Vol. 24. pp. 91–159). San Diego: Academic Press. Sorich, L., & Dweck, C.S. (1996). [Lay theories as predictors of attributions for and response to wrongdoing]. Unpublished raw data. Sorrentino, R.M., & Roney, C.J.R. (2000). The uncertain mind: Individual differences in facing the unknown. Philadelphia: Psychology Press. Sorrentino, R.M., & Short, J.C. (1986). Uncertainty orientation, motivation, and cognition. In R.M.Sorrentino & E.T. Higgins (Eds.), Handbook of motivation and cognition (pp. 379–403). New York: Guilford Press. Spears, R. (1995). Isolating the collective self. In A.Oosterwegel & R.Wicklund (Eds.), The self in European and North American culture: Development and processes (Nato ASI series, Vol. 84, pp. 309–322). Amsterdam: Kluwer. Spears, R. (2001). The interaction between the individual and the collective self: Self-categorization in context. In C.Sedikides and M.B.Brewer (Eds), Individual self, relational self, collective self (pp. 171–198). Philadelphia: Psychology Press. Spears, R. (2002). Five degrees of stereotype formation: Differentiation by any means necessary’. In C.McGarty, V.Y.Yzerbyt, & R.Spears (Eds.), Stereotypes as explanations: The formation of meaningful beliefs about social groups (pp. 127–156).Cambridge: Cambridge University Press. Spears, R., Doosje, B.J., & Ellemers, N. (1997). Self-stereotyping in the face of threats to group status and distinctiveness: The role of group identification. Personality and Social Psychology Bulletin, 23, 538– 553. Spears, R., Doosje, B., & Ellemers, N. (1999). Commitment and the context of social perception. In N.Ellemers, R.Spears, & B. Doosje (Eds.), Social identity: Context, commitment, content (pp. 59–83). Oxford: Blackwell.
364
REFERENCES
Spears, R., & Haslam, S.A. (1997). Stereotyping and the burden of cognitive load. In R. Spears, P.J.Oakes, N.Ellemers, & S.A. Haslam (Eds.), The social psychology of stereotyping and group life (pp. 171–207). Oxford: Blackwell. Spears, R., Jetten, J., Arend, S., Van Norren, M., & Postmes, T. (in prep.). Meaning and differentiation in minimal groups: Creative and reflective distinctiveness. Manuscript in preparation, University of Amsterdam/University of Queensland. Spears, R., Jetten, J., & Doosje, B. (2001). The (il)legitimacy of ingroup bias: From social reality to social resistance. In J.Jost and B. Major (Eds), The psychology of legitimacy: Emerging perspectives on ideology, justice, and intergroup relations (pp. 332–362). New York: Cambridge University Press. Spears, R., Jetten, J., & Scheepers, D. (2002). Distinctiveness and the definition of collec tive self. In A.Tesser, J.V.Wood, & D.A. Stapel (Eds.), Psychological perspectives on self and identity (Vol. 2, pp. 147–171). Lexington: American Psychological Association. Spears, R., & Lea, M. (1992). Social influence and the influence of the “social” in computer-mediated communication. In M.Lea (Ed.), Contexts of computer-mediated communication (pp. 30–65). Hemel-Hempstead, UK: Harvester Wheatsheaf. Spears, R., & Lea, M. (1994). Panacea or panopticon: The hidden power in computer-mediated communication. Communication Research, 21, 427–459. Spears, R., Lea, M., & Lee, S. (1990). De-individuation and group polarization in computer-mediated communication. British Journal of Social Psychology, 29, 121–134. Spears, R., Oakes, P.J., Ellemers, N., & Haslam, S.A. (Eds.) (1997). The social psychology of stereotyping and group life. Oxford: Blackwell. Spears, R., & Postmes, T. (1998, October). Anonymity in computer-mediated communication: Different groups have different SIDE effects. Paper presented at Society of Experimental Social Psychology meeting, Lexington, Kentucky. Srull, T.K. (1981). Person memory: Some tests of associative storage and retrieval models. Journal of Experimental Psychology: Human Learning and Memory, 7, 440–463. Stangor, C., Lynch, L., Duan, C., & Glass, B. (1992). Categorization of individuals on the basis of multiple social features. Journal of Personality and Social Psychology, 62, 207– 218. Stangor, C., & McMillan, D. (1992). Memory for expectancy-congruent and expectancy-incongruent information: A review of the social and social developmental literatures. Psychological Bulletin, 111, 42–61. Stangor, C., & Ruble, D. (1989). Strength of expectancies and memory for social information: What we remember depends on how much we know. Journal of Experimental Social Psychology, 25, 18–35. Stasser, G., Taylor, L.A., & Hanna, C. (1989). Information sampling in structured and unstructured discussions of threeand six-person groups. Journal of Personality and Social Psychology, 57, 67–78. Stasser, G., & Titus, W. (1985). Pooling of unshared information in group decision making: Biased information sampling during discussion. Journal of Personality and Social Psychology, 48, 1467–1478. Stasser, G., & Titus, W. (1987). Effects of information load and percentage of shared information on the dissemination of unshared information during group discussion. Journal of Personality and Social Psychology, 53, 81–93. Staub, E. (1989). The roots of evil: The psychological and cultural origins of genocide and other forms of group violence. New York: Cambridge University Press. Staw, B.M. (1997). The escalation of commitment: An update and appraisal. In Z.Shapira (Ed.), Organizational decision making (pp. 191–215). New York: Cambridge University Press. Staw, B.M., McKechnie, P.I., & Puffer, S.M. (1983). The justification of organizational performance. Administrative Science Quarterly, 28, 582–600. Staw, B.M., Sandelands, L.E., & Dutton, J. E. (1981). Threat-rigidity effects in organizational behavior: A multilevel analysis. Administrative Science Quarterly, 26, 501–524. Steele, C. (1997). Threat in the air: How stereotypes shape intellectual identity and performance. American Psychologist, 52, 613– 629 Stein, E. (Ed.). (1990). Forms of desire: Sexual orientation and the social constructionist controversy. New York: Garland.
REFERENCES
365
Stephan, W.G. (1977). Cognitive differentiation in intergroup perception. Sociometry, 40, 50–58. Sternberg, R.J. (1985). Lay theories of intelligence, creativity, and wisdom. Journal of Personality and Social Psychology, 49, 607– 627. Stevens, L., & Fiske, S. (2000). Motivated impressions of a powerholder: Accuracy under task dependency and misperception under evaluative dependency. Personality and Social Psychology Bulletin, 26, 907–922. Stewart, T.L., & Vassar, P.M. (2000). The effect of occupational status cues on memory for male and female targets. Psychology of Women Quarterly, 24, 161–169. Stewart, T.L., Vassar, P.M., Sanchez, D.T., & David, S.E. (2000). Attitude toward women’s societal roles moderates the effect of gender cues on target individuation. Journal of Personality and Social Psychology, 79, 143–157. Stott, C., & Drury, J. (2000). Crowds, context and identity: Dynamic categorization processes in the “poll tax riot.” Human Relations, 53, 247–273. Stroebe, K., Lodewijkx, H.F.M., & Spears, R. (2003). Do unto others as they do unto you: Bounded reciprocation versus identification as determinants of ingroup favoritism. Manuscript submitted for publication, University of Utrecht and University of Amsterdam. Su, S.K., Chiu, C., Hong, Y., Leung, K.Peng, K., & Morris, M.W. (1999). Self-organization and social organization: U.S. and Chinese constructions. In T.R.Tyler, R.M. Kramer, & O.P.John (Eds.), The psychology of the social self (pp. 193–222). Mahwah, NJ: Erlbaum. Surra, C.A. (1988). The influence of the interactive network on developing relationships. In R.M.Milardo (Ed.), Families and social networks (pp. 48–82). Beverly Hills, CA: Sage. Susskind, J., Maurer, K., Thakkar, V., Hamilton, D.L., & Sherman, J.W. (1999). Perceiving individuals and groups: Expectancies, dispositional inferences, and causal attributions. Journal of Personality and Social Psychology, 76, 181–191. Sutton, R.I. (1987). The process of organizational death: Disbanding and reconnecting. Administrative Science Quarterly, 32, 542– 569. Swim, J.K., Aiken, K.J., Hall, W.S., & Hunter, B.A. (1995). Sexism and racism: Old-fashioned and modern prejudices. Journal of Personality and Social Psychology, 68, 199– 214. Tajfel, H. (1959). Quantitative judgment in social perception. British Journal of Psychology, 50, 16–29. Tajfel, H. (1969). Cognitive aspects of prejudice. Journal of Social Issues, 25, 79–97. Tajfel, H. (1970) Experiments in intergroup discrimination. Scientific American, 233(5), 96–102. Tajfel, H. (1972). Social categorization. English manuscript of “La catégorisation sociale.” In S.Moscovici (Ed.), Introduction à la psychologic sociale (Vol. 1, pp. 272–302). Paris: Larousse. Tajfel, H. (1974). Social identity and intergroup behaviour. Social Science Information, 13, 65–93. Tajfel, H. (1978a). Differentiation between social groups. San Diego, CA: Academic Press. Tajfel, H. (1978b). Social categorization, social identity, and social comparison. In Tajfel, H. (Ed.), Differentiation between social groups: Studies in the social psychology of intergroup relations (pp. 61–76). London: Academic Press. Tajfel, H. (1978c). The social psychology of minorities. London: Minority Rights Group. Tajfel, H. (1981a). Social stereotypes and social groups. In J.C.Turner & H.Giles (Eds.), Intergroup behaviour (pp. 144–167). Oxford: Blackwell. Tajfel, H. (1981b). Human groups and social categories. Cambridge: Cambridge University Press. Tajfel, H. (1982). Social psychology- of intergroup relations. Annual Review of Psychology, 33, 1–39. Tajfel, H., Flament, C., Billig, M.G., & Bundy, R.P. (1971). Social categorization and intergroup behaviour. European Journal of Social Psychology, 1, 149–178. Tajfel, H., & Turner, J.C. (1979) An integrative theory of intergroup conflict. In W.G. Austin & S.Worschel (Eds.), The social psychology of intergroup relations (pp. 33–47). Monterey, CA: Brooks/Cole. Tajfel, H., & Turner, J.C. (1986). The social identity theory of intergroup behavior. In S. Worchel & W.G.Austin (Eds.), Psychology of intergroup relations (pp. 7–24). Chicago: Nelson-Hall. Tajfel, H., & Wilkes, A.L. (1963). Classification and quantitative judgment. British Journal of Psychology, 54, 101–114.
366
REFERENCES
Taylor, D.M., & McKirnan, D.J. (1984). A five-stage model of intergroup relations. British Journal of Social Psychology, 23, 291–300. Taylor, S.E. (1981). A categorization approach to stereotyping. In D.L.Hamilton (Ed.), Cognitive processes in stereotyping and intergroup behavior (pp. 83–114). Hillsdale, NJ: Erlbaum. Taylor, S.E. (1998). The social being in social psychology. In D.T.Gilbert, S.T.Fiske, & G.Lindzey (Eds.), The handbook of social psychology (4th ed., vol. 1, pp. 58–95). New York: McGraw-Hill. Taylor, S.E., Fiske, S.T., Etcoff, N.J., & Ruderman, A.J. (1978). Categorical and contextual bases of person memory and stereotyping. Journal of Personality and Social Psychology, 36, 778–793. Thakkar, V. (2003). The role of entitativity in judgments about groups. Unpublished doctoral disseration, University of California, Santa Barbara. Thibaut, J.W., & Kelley, H.H. (1959). The social psychology of groups. New York: Wiley. Thomason, B.C. (1982). Making sense of reification: Alfred Schutz and construction-ist theory. Atlantic Highlands, NJ: Humanities Press. Thompson, M.M., & Zanna, M.P. (1995). The conflict individual: Personality-based and domain-specific antecedents of ambivalent social attitudes. Journal of Personality, 63, 259–288. Thompson, M.S., Judd, C.M., & Park, B. (2000). The consequences of communicating social stereotypes. Journal of Experimental Social Psychology, 36, 567–599. Tice D.M. (1992). Self-presentation and self-concept change: The looking glass self as magnifying glass. Journal of Personality and Social Psychology, 63, 435–451. Tice, D.M., Butler, J.L., Muraven, M.B., & Stillwell, A.M. (1995). When modesty prevails: Differential favorability of self-presen tation to friends and strangers. Journal of Personality and Social Psychology, 69, 1120– 1138. Tönnies, F. (1955). Community and association. London: Routledge & Kegan Paul. (Originally published in German in 1887) Trafimow, D., Triandis, H.C., & Goto, S. (1991). Some tests of the distinction between private self and collective self. Journal of Personality and Social Psychology, 60, 649–655. Triandis, H.C. (1995). Individualism and collectivism. Boulder, CO: Westview. Tropp, L.R., & Wright, S.C. (2001). Ingroup identification as the inclusion of ingroup in the self. Personality and Social Psychology Bulletin, 27, 585–600. Turner, J.C. (1975). Social comparison and social identity: Some prospects for intergroup behaviour. European Journal of Social Psychology, 5, 5–34. Turner, J.C. (1982). Towards a cognitive redefinition of the social group. In H.Tajfel (Ed.), Social identity and intergroup relations (pp. 15–40). Cambridge: Cambridge University Press. Turner, J.C. (1984) Social identification and psychological group formation. In H.Tajfel (Ed.), The social dimension: European developments in social psychology (Vol. 2, pp. 518–538). Cambridge: Cambridge University Press, and Paris: Editions de la Maison des Sciences de l’Homme. Turner, J.C. (1985). Social categorization and the self-concept: A social cognitive theory of group behaviour. In E.J.Lawler (Ed.), Advances in group processes: Theory and Research (Vol. 2, pp. 77–122). Greenwich, CT: JAI. Turner, J.C. (1987a) Rediscovering the social group. In J.C.Turner, M.A.Hogg, P.J. Oakes, S.D. Reicher, & M.S. Wetherell, Rediscovering the social group: A self-categorization theory (pp. 19–41). Oxford: Blackwell. Turner, J.C. (1987b). A self-categorization theory. In J.C.Turner, M.A.Hogg, P.J. Oakes, S.D. Reicher, M.S., & Wetherel (Eds.), Rediscovering the social group: A self-categorization theory (pp. 42–67). Oxford: Blackwell. Turner, J.C. (1991). Social influence. Milton Keynes, UK: Open University Press. Turner, J.C. (1999a, July 7–11). The prejudiced personality and social change: A self-categorization perspective. The Tajfel Lecture at the general meeting of European Association of Experimental Social Psychology, Oxford. Turner, J.C. (1999b). Some current issues in research on social identity and self-categorization theories. In N.Ellemers, R.Spears, & B.Doosje (Eds.), Social identity: Context, commitment, content, (pp. 6–34). Oxford & Cambridge, MA: Blackwell. Turner, J.C. (2001). Foreword. In S.A. Haslam, Psychology in organizations: The social identity approach (x–xiii). London: Sage.
REFERENCES
367
Turner, J.C., & Brown, R.J. (1978). Social status, cognitive alternatives, and intergroup relations. In H.Tajfel (Ed.), Differentiation between social groups (pp. 201–234). London: Academic Press. Turner, J.C., Brown, R.J., & Tajfel, H. (1979). Social comparison and group interest in ingroup favouritism. European Journal of Social Psychology, 9, 187–204. Turner, J.C., Hogg, M.A., Oakes, P.J., Reicher, S.D., & Wetherell, M.S. (1987). Rediscovering the social group: A selfcategorization theory. Oxford: Blackwell. Turner, J.C., & Oakes, P.J. (1986). The significance of the social identity concept for social psychology with reference to individualism, interactionism, and social influence. British Journal of Social Psychology, 25, 237–252. Special issue on the Individual/ Society Interface. Turner, J.C., & Oakes, P.J. (1997). The socially structured mind. In C.McGarty & S. A.Haslam (Eds.), The message of social psychology (pp. 355–373). Oxford: Blackwell. Turner, J.C., Oakes, P.J., Haslam, S.A., & McGarty, C. (1994). Self and collective: Cognition and social context. Personality and Social Psychology Bulletin, 20, 454–463. Tyler, T.R., & Blader, S. (2000). Co-operation in groups: Procedural justice, social identity and behaviorl engagement. Philadelphia: Psychology Press. Unger, R.K., Draper, R.D., & Pendergrass, M.L. (1986). Personal epistemology and personal experience. Journal of Social Issues, 42, 67–79. Van Avermaet, E. (1996). Social influence in groups. In M.Hewstone, W.Stroebe, & G. Stephenson (Eds.), Introduction to social psychology: A European perspective. Oxford: Blackwell. van Knippenberg, A., van Twuyver, M., & Pepels, J. (1994). Factors affecting social categorization processes in memory. British Journal of Social Psychology, 33, 419–431. Van Schaik, C.P. (1983). Why are diurnal primates living in groups? Behaviour, 87, 120– 144. Van Twuyver, M., & van Knippenberg, A. (1998). Effects of group membership and identification on categorization and subtyping in memory. European Journal of Social Psychology, 28, 531–553. van Twuyver, M., & van Knippenberg, A. (1999). Social categorization as a function of relative group size. British Journal of Social Psychology, 38, 135–156. Verkuyten, M., & Kwa, G.A. (1996). Ethnic self-identification, ethnic involvement, and group differentiation among Chinese youth in the Netherlands. Journal of Social Psychology, 136, 35–48. Veroff, J., Sutherland, L., Chadiha, L., & Ortega, R.M. (1993). Newlyweds tell their stories: A narrative method for assessing marital experiences. Journal of Social and Personal Relationships, 10, 437–457. Vinokur, A., & Bernstein, E. (1978). Novel argumentation and attitude change: The case of polarization following group discussion. European Journal of Social Psychology, 8, 335–348. Voci, A, (2000), Perceived group variability and the salience of personal and social identity. In W.Stroebe & M.Hewstone (Eds.), European Review of Social Psychology (Vol. 11, pp. 177–221). Chichester, UK: Wiley. Voci, A., & Capozza, D. (1999, July). Outgroup homogeneity: Measures and determinants of the effect. Paper presented at the meeting of the European Association of Experimental Social Psychology, Oxford, England. Wagner, D.G., & Berger, J. (1997). Gender and interpersonal task behaviors: Status expectation accounts. Sociological Perspectives, 40, 1–32. Walster, E., Aronson, E., & Abrahams, D. (1966). On increasing the persuasiveness of a low-prestige communicator. Journal of Experimental Social Psychology, 2, 325–342. Wanous, J.P. (1976). Organizational entry: From naïve expectations to realistic beliefs. Journal of Applied Psychology, 61, 22–29. Wanous, J.P. (1977). Organizational entry: Newcomers moving from outside to inside. Psychological Bulletin, 84, 601–618. Wanous, J.P. (1980). Organizational entry: Recruitment, selection, and socialization of newcomers. Reading, MA: Addison-Wesley. Wanous, J.P. (1989). Impression management at organizational entry. In R.A.Giacalone & P.Rosenfeld (Eds.), Impression management in the organization (pp. 253–367). Hillsdale, NJ: Erlbaum.
368
REFERENCES
Wanous, J.P., Poland, T.D., Premack, S.L., & Davis, K.S. (1992). The effects of met expectations on newcomer attitudes and behaviors: A review and meta-analysis. Journal of Applied Psychology, 77, 288–297. Warriner, C.K. (1956). Groups are real: A reaffirmation. American Sociological Review, 21, 549–554. Watson, D., & Tellegen, A. (1985). Toward a consensual structure of mood. Psychological Bulletin, 98, 219–235. Weary, G., Marsh, K.L., Gleicher, F., & Edwards, J.A. (1993). Depression, control motivation, and the processing of information about others. In G.Weary, F.Gleicher, & K.L.Marsh (Eds.), Control motivation and social cognition (pp. 255–287). New York: Springer-Verlag. Weber, R., & Crocker, J. (1983). Cognitive processes in the revision of stereotypic beliefs. Journal of Personality and Social Psychology, 45, 961–977. Webster, D.M., & Kruglanski, A. (1994). Individual differences in need for cognitive closure. Journal of Personality and Social Psychology, 67, 1049–1062. Wegener, D.T., & Petty, R.E. (1998). The naive scientist revisited: Naive theories and social judgment. Social Cognition, 16, 1–7. Wegener, I. (2000). Soziale Kategorien im situativen Kontext: Kognitive Flexibilität bei der Personenwahrnehmung [Social categories in their situational context: Cognitive flexibility in person perception]. Unpublished doctoral dissertation, Universität Bonn, Germany. Weick, K.E., & Roberts, K.H. (1993). Collective mind in organizations: Heedful relating on flight decks. Administrative Science Quarterly, 38, 357–381. Weiner, B., Perry, R., & Magnusson, J. (1988). An attributional analysis of reactions to stigmas. Journal of Personality and Social Psychology, 55, 738–748. Welbourne, J.L. (1999). The impact of perceived entitativity on inconsistency resolution for groups and individuals. Journal of Experimental Social Psychology, 35, 481–508. West, T.L., & Levy, S.R. (2003). Lay theories and helping. Manuscript submitted for publication. Whitehead, A.N. (1929). Process and reality. New York: Free Press. Whitehead, A.N. (1938). Modes of thought. New York: Free Press. Wilder, D.A. (1978). Reduction of intergroup discrimination through individuation of the outgroup. Journal of Personality and Social Psychology, 36, 1361–1374. Wilder, D.A. (1981). Perceiving persons as a group: Categorization and intergroup relations. In D.L.Hamilton (Ed.), Cognitive processes in stereotyping and intergroup behaviour (pp. 213–257). Hillsdale, NJ: Erlbaum. Wilder, D.A. (1984). Predictions of belief homogeneity and similarity following social categorization. British Journal of Social Psychology, 23, 323–333. Wilder, D.A. (1986). Social categorization: Implications for creation and reduction of intergroup conflicts. In L.Berkowitz (Ed.), Advances in experimental social psychology (Vol. 19, pp. 293–355). San Diego, CA: Academic Press. Wilder, D.A. (1990). Some determinants of the persuasive power of in-groups and out-groups: Organization of information and attribution of independence. Journal of Personality and Social Psychology, 59, 1202– 1213. Wilder, D.A., & Simon, A.F. (1998). Categorical and dynamic groups: Implications for social perception and intergroup behavior. In C.Sedikides, J.Schopler, & C.A.Insko (Eds.) Intergroup cognition and intergroup behavior (pp. 27–44). New Jersey: Erlbaum. Wilder, D.A., & Thompson, J.E. (1980). In tergroup contact with independent manipu-lations on in-group and outgroup interaction. Journal of Personality and Social Psychology, 38, 589–603. Wildschut, T., Insko, C.A., & Gaertner, L. (2002). Intragroup social influence and intergroup competition. Journal of Personality and Social Psychology, 82, 975–992. Wildschut, T., Lodewijkx, H.F.M., & Insko, C.A. (2001). Toward a reconciliation of diverging perspectives on interindividual-intergroup discontinuity: The role of procedural interdependence. Journal of Experimental Social Psychology, 37, 273–285. Wildschut, T., Pinter, B., Vevea, J.L., Insko, C.A., & Schopler, J. (2003). Beyond the group mind: A quantitative review of the interindividual-intergroup discontinuity effect. Psychological Bulletin, 129, 698–722.
REFERENCES
369
Williams, J.A. (1984). Gender and intergroup behaviour: Towards an integration. British Journal of Social Psychology, 23, 311–316. Williams, K.D., Cheung, K.T., & Choi, W. (2000). Cyberostracism: Effects of being ignored over the Internet. Journal of Personality and Social Psychology, 79, 748–762. Williams, K.D., & Zadro, L. (2001). Ostracism: On being ignored, excluded, and rejected. In M.R.Leary (Ed.), Interpersonal rejection (pp. 21–53). New York: Oxford University Press. Willis, H.J., Rubin, M., & Hewstone, M. (2000, March). An alternative explanation of out-group homogeneity: Selfanchoring theory. Poster presented at the Annual Conference of the British Psychological Society, Winchester, England. Winquist, J.R., Larson, J.R., Jr. (1998). Information pooling: When it impacts group decision making. Journal of Personality and Social Psychology, 74, 371–377. Wittenbaum, G.M., & Stasser, G. (1998). The reconstrual of information during group discussion. Group Processes and Intergroup Relations, 1, 78–87. Wittenbrink, B., Hilton, J.L., & Gist, P.L. (1998). In search of similarity: Stereotypes as naive theories in social categorization. Social Cognition, 16, 31–55. Wolf, S., & Latané, B. (1985). Conformity, innovation and the psychosocial law. In S. Moscovici, G.Mugny & E.Avermaet (Eds.), Perspectives on minority influence (pp. 201– 215). Cambridge: Cambridge University Press. Wolsko, C. Park, B., Judd, C.M., & Wittenbrink, B. (2000). Framing interethnic ideology: Effects of multicultural and colorblind perspectives on judgments of groups and individuals. Journal of Personality and Social Psychology, 78, 635–654. Wood, W., Lundgren, S., Ouellette, J.A., Busceme, S., & Blackstone, T. (1994). Minority influence: A meta-analytic review of social influence processes. Psychological Bulletin, 115, 323–345. Worchel, S., Coutant-Sassic, D., & Grossman, M. (1992). A developmental approach to group dynamics: A model and illustrative research. In S.Worchel, W.Wood, & J.A. Simpson (Eds.), Group process and productivity (pp. 181–202). Newbury Park, CA: Sage. Wrangham, R.W. (1979). On the evolution of ape social systems. Social Science Information, 18, 334–368. Wright, S.C., Taylor, D.M., & Moghaddam, F.M. (1990) Responding to membership in a disadvantaged group: From acceptance to collective protest. Journal of Personality and Social Psychology, 58, 994–1003. Yamagishi, T. (2001, July) Exchange and solidarity mechanisms: Evolutionary basis of ingroup attachment and ingroup favoring mechanisms. Paper presented at Seventh Annual SASP meeting, Symposium on “Asian Research on Social Identity.” Melbourne, Australia. Young, H., van Knippenberg, A., Ellemers, N., & de Vries, N. (1997). The effects of group membership and social context on information organization. European Journal of Social Psychology, 27, 523–537. Young, H., van Knippenberg, A., Ellemers, N., & de Vries, N. (1999). The asymmetrical perception of men and women. Group Processes and Intergroup Relations, 2, 259–278. Yzerbyt, V.Y., & Buidin, G. (1998). The impact of naive theories about the nature of the groups on the accentuation effect. Unpublished raw data, Catholic University of Louvain at Louvain-la-Neuve, Belgium. Yzerbyt, V.Y., Castano, E., Leyens, J-Ph., & Paladino, M.P. (2000). The primacy of the ingroup: The interplay of entitativity and identification. In W.Stroebe & M.Hewstone (Eds.), European review of social psychology (Vol. 11, pp. 257–275). Chichester, UK: Wiley Yzerbyt, V.Y., Corneille, O., & Estrada, C. (2001). The interplay of naive theories and entitativity from the outsider and the insider perspectives. Personality and Social Psychology Review, 5, 141–155. Yzerbyt, V. Y., Coull, A., & Rocher, S.J. (1999). Fencing off the deviant: The role of cognitive resources in the maintenance of stereotypes. Journal of Personality and Social Psychology, 77, 449–462. Yzerbyt, V.Y., & Estrada, C. (1999). The role of subjective essentialism on group decision making. Unpublished raw data. Catholic University of Louvain at Louvain-la-Neuve, Belgium. Yzerbyt, V.Y., Leyens, J.-P., & Bellour, F. (1995). The ingroup overexclusion effect: Identity concerns in decisions about group membership. European Journal of Social Psychology, 25, 1–16.
370
REFERENCES
Yzerbyt, V.Y., Leyens, J.P., & Corneille, O. (1998). Social judgeability and the bogus pipeline: The role of naive theories of judgment in impression formation. Social Cognition, 16, 56–77. Yzerbyt, V Y., & Rocher, S. (2002). Subjective essentialism and the emergence of stereotypes. In C.McGarty, V.Y.Yzerbyt & R. Spears (Eds.), Stereotypes as explanations: The formation of meaningful beliefs about social groups (pp. 38–66). Cambridge: Cambridge University Press. Yzerbyt, V.Y., Rocher, S., McGarty, C., & Haslam, S.A. (1997). Illusory correlation: Naive theories and the formation of social stereotypes. Unpublished raw data. Catholic University of Louvain at Louvain-laNeuve, Belgium. Yzerbyt, V.Y., Rocher, S., & Schadron, G. (1997). Stereotypes as explanations: A subjective essentialistic view of group perception. In R.Spears, P.Oakes, N.Ellemers, & S.A.Haslam (Eds.), The psychology of stereotyping and group life (pp. 20–50). Oxford: Blackwell. Yzerbyt, V.Y., & Rogier, A. (2001). Blame it on the group: Entitativity, subjective essen tialism and social attribution. In J.Jost & B. Major (Eds.), The psychology of legitimacy: Emerging perspectives on ideology, justice, and intergroup relations (pp. 103–134). New York: Cambridge University Press. Yzerbyt, V.Y., Rogier, A., & Fiske, S. (1998). Group entitativity and social attribution: On translating situational constraints into stereotypes. Personality and Social Psychology Bulletin, 24, 1090–1104. Zajonc, R.B., (1960). The process of cognitive tuning in communication. Journal of Abnormal and Social Psychology, 61, 159–167. Zander, A. (1976). The psychology of removing group members and recruiting new ones. Human Relations, 29, 969–987. Zdaniuk, B., & Levine, J. (1996). Anticipated interaction and thought generation: The role of faction size. British Journal of Social Psychology, 35, 201–218.
Index
agency, 4–6, 27–31, 33, 36–38, 58, 89, 90, 93, 94, 96, 97, 100, 267, 268, 272, 399 attributional process, 7 availability heuristic, 165
comparative fit, 16, 152, 153, 160, 299, 321 competition, 18, 87, 306–308, 310, 327, 343, 344, 351– 355, 358, 359 covariation, 12, 13, 31, 150, 152, 154, 155, 179–202
between-category confusions, 11, 150 black sheep effect, 18, 367, 368, 370, 379
discontinuity effect, 17, 336, 339, 345, 348, 349, 351–353 discrimination, 7, 18, 21, 22, 103, 122, 123, 156, 201, 221, 235, 297, 300, 307, 319, 333, 398, 404, 408 dispersion, 13, 15, 134, 147, 203, 207, 212–215, 217, 219, 243–251, 254, 284–286, 288–290, 294, 320, 331 distinctiveness, 16, 18, 114, 117, 121, 155, 156, 165, 166, 170–172, 222, 293–309, 311, 313, 315, 327, 363, 368, 374, 386, 395, 405, 409, 410, 412 diversity, 2, 3, 6, 27, 98, 100, 169, 173, 178, 184, 255, 284, 401, 412
categorization, 2, 10, 11, 39, 40, 42, 48, 63, 65, 66, 68, 111, 121, 134, 145, 147, 148, 151–153, 156, 157, 161– 164, 168, 175, 181, 204, 300, 303, 305, 312, 318, 319, 328, 331, 358, 359, 362–364, 366, 373, 398, 408, 413, 414 category boundaries, 63, 76, 81, 88, 95, 97, 98, 148 coherence, 9, 10, 16, 18, 19, 26, 36, 56, 65, 66, 68, 69, 89, 110, 111, 128, 142, 143, 168, 174, 176, 177 cohesiveness, 10, 16, 307 common fate, 18, 19, 25–27, 33, 45, 55, 89, 90, 104, 107, 108, 122, 200, 264–266, 268, 317–319, 355, 362, 382– 384, 386–390, 412, 417, 420, 421, 435 common goals, 4, 5, 27, 29, 45, 46, 55, 56, 89, 90, 94, 97, 104, 107, 108, 190, 191, 294, 316, 362. communication, 10, 14, 15, 17, 31, 91, 106, 227, 228, 231, 236, 237–240, 244, 247, 249, 250, 253–255, 257, 258, 260–263, 268–273, 305, 340, 343, 344, 390, 407
entitativity, 1–10, 12–22, 25–40, 43–49, 51–61, 65–67, 69–79, 81, 88–100, 102–109, 111, 116, 121, 122, 127– 129, 132, 138, 140–143, 161, 168, 169, 173, 175–177, 190, 191, 194, 197, 200, 201, 253, 258, 263–273, 293– 299, 301–309, 311, 313–315, 317–321, 328, 329, 331, 333, 349, 355, 356, 361–363, 375, 378, 379, 381–396, 398, 399, 401, 402, 409–412, 414–422, 428, 429, 434– 437 entiteists, 9 371
372
INDEX
essence, 2–4, 6, 7, 9, 14, 15, 27–29, 33, 35–38, 42, 44, 52, 53, 58–60, 62, 63, 65, 67, 68, 70, 75–77, 90–95, 97, 98, 100, 102, 103, 105, 106, 108, 109, 112, 114, 115, 117, 122–124, 127, 130, 141–143, 175, 266, 268, 294, 315, 410 essentialism, 1, 3–8, 14, 21, 22, 27, 38–40, 42, 45, 49, 51– 53, 57–59, 61–65, 67–69, 74–78, 95, 101–106, 109, 111–113, 115, 116, 120–124, 161, 258, 265, 270, 272, 362, 418 exemplar, 80, 83, 85, 191, 226, 227, 254, 357, 410 expectations, 11, 12, 17, 114, 127–129, 141, 166, 167, 169, 171, 173, 174, 177, 237, 280–282, 284, 304, 321, 322, 337, 341–343, 347–349, 351, 357, 364, 365, 426 familiarity, 10, 12, 13, 180, 181, 183, 184, 187–189, 191– 194, 196, 197, 199, 200, 203–206, 212–214, 216, 220, 277, 278, 435 fit, 16, 80, 83–85, 88, 151–154, 159, 160, 164, 232, 282, 286, 297, 299, 321, 322, 324, 362, 364, 378, 388, 390, 393, 415 group boundaries, 2, 5, 7, 17, 29, 30, 79–81, 84, 86–88, 90, 95, 97, 98, 100, 104, 122, 235, 267, 312, 329, 330, 332 groupness, 6, 7, 19, 20, 30, 33, 40, 43, 45–47, 49, 51, 53, 55, 57, 59, 66, 94, 119, 120, 141, 142, 294, 299, 349, 355 homogeneity, 2, 4–10, 11, 13–18, 26, 39–41, 44, 46, 48, 49, 52, 53, 55–59, 65, 67–69, 72, 74–77, 81, 82, 90, 93, 94, 96–100, 102, 105, 107, 109, 110, 112, 116, 118, 127–129, 134, 137, 138, 140–143, 147–151, 153–156, 158–160, 180, 181, 184, 185, 188, 191, 197–201, 203, 204, 208, 209, 211, 218, 221, 223, 229, 234, 278, 279, 285, 294–297, 309–311, 313, 314, 320, 321, 328, 331, 332, 356, 379, 397 human artifacts, 2, 42, 64, 91 illusory correlation, 11, 112, 146, 165–167, 169–174, 177 immutability, 43, 67, 68, 71, 72, 74–76, 95, 100, 104, 146, 266, 412 incrementalists, 9 inductive potential, 5, 42–44, 52, 59, 63, 67, 68, 71, 72, 74, 75, 91, 95, 97, 100, 105, 106, 266, 272 intergroup conflict, 8, 91, 221, 235, 352, 356, 398 intimacy groups, 2, 4, 7, 29, 46, 47, 49, 51, 55, 176, 272 lay theories, 9, 127–130, 132, 135, 137, 138, 141, 142, 145
legitimacy, 324, 326, 398, 405 majority-minority, 221, 222, 225 meta-contrast ratio, 376 mixed-motive game, 17 natural kinds, 2, 3, 5, 27, 42, 52, 63, 67, 69, 72, 75, 76, 90, 91, 95, 97, 104, 175, 267, 268, 294 normative fit, 16, 17, 321, 322, 362, 364 outcome dependency, 221, 222, 224, 225, 227–229, 234, 235 outgroup fear, 18, 341, 351–353, 355–358 outgroup homogeneity effect, 2, 13, 41, 128, 180, 221, 229, 234, 279, 285, 310 outgroup schema, 17, 18, 335–337, 339, 341, 342, 347– 349, 351–357 permeability, 9, 17, 46, 90, 95, 100, 267, 312, 313, 324, 329, 330, 332, 405 phenotypic attribute, 6 power, 2, 13, 17, 19, 31, 72, 81, 111, 137, 221–232, 234– 236, 304, 319, 320, 325–332, 363, 392, 396, 412, 417, 436 prejudice, 7, 21, 64, 70, 71, 77, 129, 130, 131, 136, 143, 144, 146, 240, 319, 325, 379, 408 procedural interdependence, 18, 348–350, 352–356 prototype, 4, 20, 30, 44, 63, 362, 404–406, 408, 410–412, 414, 416, 417 rumor transmission, 14, 239, 242 self-categorization, 16, 39, 48, 159, 160, 162–164, 166, 178, 278, 296, 297, 301, 303, 312, 315, 319–323, 325, 332, 333, 362–364, 385, 405, 406, 408–412, 414, 417 similarity, 4–8, 10, 13, 19, 25–30, 33, 44–46, 48, 53–59, 65, 67, 76, 89, 90, 91, 93–98, 102–105, 107–110, 112, 115–117, 119, 122, 133, 135, 142, 145, 147–150, 153, 156, 163, 164, 169, 190, 200, 204, 212, 217, 265–267, 269, 270, 294, 305, 313, 314, 317, 318, 355, 356, 362, 372, 374, 375, 383, 384, 386, 388–391, 412, 417, 420, 435 size, 2, 11, 13, 29, 33, 46, 47, 128, 149, 151, 155, 156, 159, 168–170, 203–205, 210–217, 220–227, 229, 230, 232–235, 265, 296, 362, 429 social categories, 2–4, 7, 8, 10, 12, 14, 29, 30, 34, 42, 43, 45–53, 55, 57–62, 65–71, 73–80, 90, 91, 95, 98, 104, 109, 124, 148, 175–179, 182, 187, 191, 193, 197, 200,
INDEX
201, 217, 253, 255, 258, 265, 267, 268, 270, 271, 362, 366, 386 social identification, 17, 19, 21, 22, 38, 163, 192, 226, 382, 399, 402, 405, 406, 408, 417 social identity, 9, 16, 29, 39, 66, 159, 160, 193, 194, 208, 210–212, 217–219, 222, 264, 278, 279, 293–299, 301, 303, 304, 308–310, 313, 315, 317, 319–324, 327–329, 331, 338, 362, 363, 367, 368, 373, 377, 395, 396, 398, 402, 404–406, 408, 409, 417, 418 social integration, 20, 419–422, 424, 426, 429, 430, 435, 436 socialization, 15, 21, 26, 38, 106, 124, 131, 275, 276, 279–286, 290, 291, 425, 426 stability, 9, 43, 95, 313, 323, 324, 332, 405, 412 status, 2, 13, 16, 17, 25, 26, 28, 31, 41, 66–69, 76, 79, 91, 96, 100, 115, 123, 149, 151, 152, 154, 181, 191, 197, 204, 207–212, 215, 217–219, 221–225, 227, 229–232, 234–236, 259, 276, 280, 281, 284, 297, 298, 310–314, 317, 323–327, 329–333, 368, 377, 395, 403, 405, 406, 409, 431, 434 stereotype change, 5, 79, 81, 88, 96, 97, 291, 379 stereotypes, 2, 4, 5, 7, 9, 11, 14, 28, 30, 39, 40, 42–44, 48– 53, 57, 77, 80, 81, 84, 99, 127, 130, 132, 134–137, 140, 141, 143–145, 150, 160, 161, 163–165, 171, 172, 175, 177, 184, 232, 237–241, 244, 245, 251, 253–255, 268, 284, 285, 291, 319, 320, 325, 328–330, 332, 333, 364 stereotypicality, 134, 139, 203, 217, 218, 242–252, 284– 286, 288, 289 subgrouping, 82–84, 88, 100 subjective group dynamics, 18, 361, 363, 366, 367, 378, 379 subtyping, 82–85, 88 task groups, 2, 4, 7, 29, 46, 47, 49–51, 55, 57–59, 176, 305 terror management theory, 19, 392, 396 totalist groups, 19, 20, 22, 401–404 typicality, 83, 342, 375–377 unalterability, 42, 43, 52, 91, 95, 98, 106, 266, 267, 270, 272 uncertainty, 15, 19, 20, 45, 275, 276, 280–284, 286, 288– 290, 292, 302, 364, 401–404, 406–418 variability, 1–7, 9, 12, 13, 15, 21, 22, 29, 36, 40, 41, 46, 48, 54, 82, 83, 97, 111, 127, 128, 132–135, 137–141, 143, 144, 147, 169, 171–173, 177, 179–181, 183, 188, 190, 192, 194–196, 197–201, 205, 207–209, 211, 215, 218, 219, 221–232, 234–236, 238, 240, 242, 250, 254,
373
275–280, 282–285, 287, 288, 290, 291, 294, 296, 309– 312, 320, 321, 323, 331, 379, 412, 415 “Who said What?” paradigm, 10, 11, 148–150, 152, 154, 157, 158, 188 within-category confusions, 11, 148, 150, 154
